{"checked_at":"2026-09-15T21:39:28.377523+00:00","attribution":{"name":"Epoch AI — Capabilities & Benchmarking","url":"https://epoch.ai/benchmarks","licence":"CC BY 4.0","licence_url":"https://creativecommons.org/licenses/by/4.0/","citation":"Epoch AI, 'Capabilities & Benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks' [online resource]."},"frontier":{"id":"open-weights-gap","name":"Open weights vs the frontier","leader":{"model":"GPT-6 Astra","score":166.31,"ci_low":163.0,"ci_high":171.88,"date":"2026-09-03","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":1,"training_compute_flop":null,"frontier_days":12,"lag_months":0.0,"lag_low":0.0,"lag_high":2.8,"frontier_date":"2026-09-03","frontier_model":"GPT-6 Astra","at_frontier":true,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"best_open":{"model":"Kimi K3","score":157.63,"ci_low":155.38,"ci_high":160.45,"date":"2026-07-16","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["China"],"rank":11,"training_compute_flop":2.0001e+25,"frontier_days":0,"lag_months":4.4,"lag_low":2.8,"lag_high":7.1,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":false,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},"gap_months":4.4,"matches":"GPT-5.4 Pro","matched_on":"2026-03-05","history":[{"date":"2023-07-18","score":113.63,"model":"Llama 2-70B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":4.1},{"date":"2023-07-20","score":117.0,"model":"Stable Beluga 2","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":4.2},{"date":"2023-11-02","score":117.3,"model":"Yi-34B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":7.7},{"date":"2023-12-11","score":118.4,"model":"Mixtral 8x7B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":8.9},{"date":"2024-04-17","score":122.0,"model":"Mixtral 8x22B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":13.1},{"date":"2024-04-18","score":122.91,"model":"Llama 3-70B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":13.2},{"date":"2024-05-07","score":124.77,"model":"DeepSeek-V2 (MoE-236B, May 2024)","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":13.8},{"date":"2024-06-07","score":125.27,"model":"Qwen2-72B","closed_first_reached":"2023-03-14","matches":"GPT-4 (Mar 2023)","gap_months":14.8},{"date":"2024-07-23","score":128.75,"model":"Llama 3.1-405B","closed_first_reached":"2024-05-13","matches":"GPT-4o (May 2024)","gap_months":2.3},{"date":"2024-09-19","score":129.0,"model":"Qwen2.5-72B","closed_first_reached":"2024-06-20","matches":"Claude 3.5 Sonnet","gap_months":3.0},{"date":"2024-12-12","score":130.42,"model":"Phi-4","closed_first_reached":"2024-09-12","matches":"o1-mini","gap_months":3.0},{"date":"2024-12-26","score":132.35,"model":"DeepSeek-V3","closed_first_reached":"2024-09-12","matches":"o1-mini","gap_months":3.4},{"date":"2025-01-20","score":138.97,"model":"DeepSeek-R1","closed_first_reached":"2024-12-17","matches":"o1","gap_months":1.1},{"date":"2025-04-28","score":139.35,"model":"Qwen3-235B-A22B","closed_first_reached":"2024-12-17","matches":"o1","gap_months":4.3},{"date":"2025-05-28","score":141.29,"model":"DeepSeek-R1 (May 2025)","closed_first_reached":"2024-12-17","matches":"o1","gap_months":5.3},{"date":"2025-07-25","score":143.88,"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","closed_first_reached":"2025-03-31","matches":"Gemini 2.5 Pro (Mar 2025)","gap_months":3.8},{"date":"2025-09-29","score":145.02,"model":"DeepSeek-V3.2-Exp","closed_first_reached":"2025-04-16","matches":"o3","gap_months":5.5},{"date":"2025-11-06","score":145.76,"model":"Kimi K2 Thinking","closed_first_reached":"2025-04-16","matches":"o3","gap_months":6.7},{"date":"2025-12-01","score":146.19,"model":"DeepSeek-V3.2","closed_first_reached":"2025-04-16","matches":"o3","gap_months":7.5},{"date":"2026-01-27","score":148.01,"model":"Kimi K2.5","closed_first_reached":"2025-08-07","matches":"GPT-5","gap_months":5.7},{"date":"2026-04-07","score":149.71,"model":"GLM-5.1","closed_first_reached":"2025-08-07","matches":"GPT-5","gap_months":8.0},{"date":"2026-04-20","score":151.0,"model":"Kimi K2.6","closed_first_reached":"2025-11-18","matches":"Gemini 3 Pro","gap_months":5.0},{"date":"2026-06-16","score":151.86,"model":"GLM-5.2","closed_first_reached":"2025-11-18","matches":"Gemini 3 Pro","gap_months":6.9},{"date":"2026-07-16","score":157.63,"model":"Kimi K3","closed_first_reached":"2026-03-05","matches":"GPT-5.4 Pro","gap_months":4.4}],"open_compute":[{"model":"LLaMA-65B","date":"2023-02-24","score":109.94,"organization":"Meta AI","lab":null,"model_id":null,"flop":5.5e+23},{"model":"LLaMA-33B","date":"2023-02-24","score":107.14,"organization":"Meta AI","lab":null,"model_id":null,"flop":2.7300000000001e+23},{"model":"LLaMA-13B","date":"2023-02-24","score":100.2,"organization":"Meta AI","lab":null,"model_id":null,"flop":7.8e+22},{"model":"LLaMA-7B","date":"2023-02-24","score":96.25,"organization":"Meta AI","lab":null,"model_id":null,"flop":4.00000001e+22},{"model":"Cerebras-GPT-13B","date":"2023-03-20","score":82.57,"organization":"Cerebras Systems","lab":null,"model_id":null,"flop":2.3e+22},{"model":"Falcon-7B","date":"2023-04-24","score":94.63,"organization":"Technology Innovation Institute","lab":null,"model_id":null,"flop":6.3e+22},{"model":"MPT-7B","date":"2023-05-05","score":94.11,"organization":"MosaicML","lab":null,"model_id":null,"flop":4.2000000000000004e+22},{"model":"Falcon-40B","date":"2023-05-25","score":104.13,"organization":"Technology Innovation Institute","lab":null,"model_id":null,"flop":2.4e+23},{"model":"Baichuan1-7B","date":"2023-06-01","score":89.92,"organization":"Baichuan","lab":null,"model_id":null,"flop":5.04e+22},{"model":"MPT-30B","date":"2023-06-22","score":100.28,"organization":"MosaicML","lab":null,"model_id":null,"flop":1.8900000000001e+23},{"model":"XGen-7B","date":"2023-06-27","score":92.9,"organization":"Salesforce","lab":null,"model_id":null,"flop":8.02e+22},{"model":"Llama 2-70B","date":"2023-07-18","score":113.63,"organization":"Meta AI","lab":null,"model_id":null,"flop":8.1e+23},{"model":"Llama 2-13B","date":"2023-07-18","score":105.88,"organization":"Meta AI","lab":null,"model_id":null,"flop":1.6e+23},{"model":"Llama 2-7B","date":"2023-07-18","score":98.66,"organization":"Meta AI","lab":null,"model_id":null,"flop":8.4e+22},{"model":"Falcon-180B","date":"2023-09-06","score":111.94,"organization":"Technology Innovation Institute","lab":null,"model_id":null,"flop":3.76e+24},{"model":"Baichuan2-13B","date":"2023-09-06","score":102.86,"organization":"Baichuan","lab":null,"model_id":null,"flop":2.03e+23},{"model":"Phi-1.5","date":"2023-09-11","score":90.99,"organization":"Microsoft","lab":null,"model_id":null,"flop":1.17e+21},{"model":"Baichuan 2-7B","date":"2023-09-20","score":95.88,"organization":"Baichuan","lab":null,"model_id":null,"flop":1.092e+23},{"model":"Qwen-14B","date":"2023-09-24","score":112.85,"organization":"Alibaba","lab":null,"model_id":null,"flop":2.5e+23},{"model":"Qwen-7B","date":"2023-09-28","score":106.55,"organization":"Alibaba","lab":null,"model_id":null,"flop":1.01e+23},{"model":"Yi-34B","date":"2023-11-02","score":117.3,"organization":"01.AI","lab":null,"model_id":null,"flop":6.1e+23},{"model":"DeepSeek Coder 33B","date":"2023-11-02","score":95.87,"organization":"DeepSeek,Peking University","lab":null,"model_id":null,"flop":3.96e+23},{"model":"DeepSeek Coder 6.7B","date":"2023-11-02","score":89.05,"organization":"DeepSeek,Peking University","lab":null,"model_id":null,"flop":8.04e+22},{"model":"DeepSeek Coder 1.3B","date":"2023-11-02","score":62.54,"organization":"DeepSeek,Peking University","lab":null,"model_id":null,"flop":1.56e+22},{"model":"Yi 6B","date":"2023-11-22","score":104.46,"organization":"01.AI","lab":null,"model_id":null,"flop":1.26e+23},{"model":"DeepSeek LLM 67B","date":"2023-11-29","score":110.56,"organization":"DeepSeek","lab":null,"model_id":null,"flop":8.04e+23},{"model":"Mixtral 8x7B","date":"2023-12-11","score":118.4,"organization":"Mistral AI","lab":null,"model_id":null,"flop":7.74e+23},{"model":"Phi-2","date":"2023-12-12","score":107.69,"organization":"Microsoft","lab":null,"model_id":null,"flop":2.27e+22},{"model":"StarCoder 2 15B","date":"2024-02-20","score":104.7,"organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","lab":null,"model_id":null,"flop":3.87e+23},{"model":"StarCoder 2 7B","date":"2024-02-20","score":93.06,"organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","lab":null,"model_id":null,"flop":1.55e+23},{"model":"Gemma 7B","date":"2024-02-21","score":111.79,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":3.07e+23},{"model":"Gemma 2B","date":"2024-02-21","score":93.72,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":4.5115822e+22},{"model":"StarCoder 2 3B","date":"2024-02-22","score":88.18,"organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","lab":null,"model_id":null,"flop":5.94e+22},{"model":"Mixtral 8x22B","date":"2024-04-17","score":122.0,"organization":"Mistral AI","lab":null,"model_id":null,"flop":2.34e+24},{"model":"Llama 3-70B","date":"2024-04-18","score":122.91,"organization":"Meta AI","lab":null,"model_id":null,"flop":7.861e+24},{"model":"Llama 3-8B","date":"2024-04-18","score":116.35,"organization":"Meta AI","lab":null,"model_id":null,"flop":7.2e+23},{"model":"phi-3-small 7.4B","date":"2024-04-23","score":121.8,"organization":"Microsoft","lab":null,"model_id":null,"flop":2.1312e+23},{"model":"phi-3-medium 14B","date":"2024-04-23","score":121.19,"organization":"Microsoft","lab":null,"model_id":null,"flop":4.032e+23},{"model":"phi-3-mini 3.8B","date":"2024-04-23","score":117.25,"organization":"Microsoft","lab":null,"model_id":null,"flop":7.524e+22},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","date":"2024-05-07","score":124.77,"organization":"DeepSeek","lab":null,"model_id":null,"flop":1.02e+24},{"model":"Falcon 2 11B","date":"2024-05-09","score":109.32,"organization":"Technology Innovation Institute","lab":null,"model_id":null,"flop":3.6e+23},{"model":"Qwen2-72B","date":"2024-06-07","score":125.27,"organization":"Alibaba","lab":null,"model_id":null,"flop":3.02e+24},{"model":"Gemma 2 27B","date":"2024-06-24","score":122.05,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":2.106e+24},{"model":"Gemma 2 9B","date":"2024-06-24","score":119.79,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":4.32e+23},{"model":"Llama 3.1-405B","date":"2024-07-23","score":128.75,"organization":"Meta AI","lab":null,"model_id":null,"flop":3.8e+25},{"model":"Llama 3.1-70B","date":"2024-07-23","score":125.91,"organization":"Meta AI","lab":null,"model_id":null,"flop":7.929e+24},{"model":"Llama 3.1-8B","date":"2024-07-23","score":116.5,"organization":"Meta AI","lab":null,"model_id":null,"flop":1.224e+24},{"model":"Qwen2.5-32B","date":"2024-09-17","score":128.52,"organization":"Alibaba","lab":null,"model_id":null,"flop":3.51e+24},{"model":"Qwen2.5-Coder-32B","date":"2024-09-18","score":119.42,"organization":"Alibaba","lab":null,"model_id":null,"flop":1.0725e+24},{"model":"Qwen2.5-Coder (7B)","date":"2024-09-18","score":112.97,"organization":"Alibaba","lab":null,"model_id":null,"flop":2.5113e+23},{"model":"Qwen2.5-Coder (1.5B)","date":"2024-09-18","score":102.66,"organization":"Alibaba","lab":null,"model_id":null,"flop":5.082e+22},{"model":"Qwen2.5-72B","date":"2024-09-19","score":129.0,"organization":"Alibaba","lab":null,"model_id":null,"flop":7.8e+24},{"model":"Qwen2.5-7B","date":"2024-09-19","score":118.44,"organization":"Alibaba","lab":null,"model_id":null,"flop":8.2188e+23},{"model":"Llama 3.2 1B","date":"2024-09-24","score":102.43,"organization":"Meta AI","lab":null,"model_id":null,"flop":6.642e+22},{"model":"Llama 3.3 70B","date":"2024-12-06","score":127.32,"organization":"Meta AI","lab":null,"model_id":null,"flop":6.8649768e+24},{"model":"Phi-4","date":"2024-12-12","score":130.42,"organization":"Microsoft Research","lab":null,"model_id":null,"flop":9.3202015e+23},{"model":"DeepSeek-V3","date":"2024-12-26","score":132.35,"organization":"DeepSeek","lab":null,"model_id":null,"flop":3.3e+24},{"model":"DeepSeek-R1","date":"2025-01-20","score":138.97,"organization":"DeepSeek","lab":null,"model_id":null,"flop":3.5e+24},{"model":"Mistral Small 3","date":"2025-01-30","score":127.07,"organization":"Mistral AI","lab":null,"model_id":null,"flop":1.152e+24},{"model":"QwQ-32B","date":"2025-03-05","score":137.6,"organization":"Alibaba","lab":null,"model_id":null,"flop":3.51e+24},{"model":"Gemma 3 27B","date":"2025-03-12","score":130.03,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":2.268e+24},{"model":"Gemma 3 12B","date":"2025-03-12","score":123.46,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":8.64e+23},{"model":"Gemma 3 4B","date":"2025-03-12","score":115.97,"organization":"Google DeepMind","lab":null,"model_id":null,"flop":9.6e+22},{"model":"DeepSeek-V3 (Mar 2025)","date":"2025-03-24","score":135.95,"organization":"DeepSeek","lab":null,"model_id":null,"flop":3.3e+24},{"model":"Llama 4 Scout","date":"2025-04-05","score":129.64,"organization":"Meta AI","lab":null,"model_id":null,"flop":4.08e+24},{"model":"Llama 4 Maverick","date":"2025-04-06","score":132.2,"organization":"Meta AI","lab":null,"model_id":null,"flop":2.244000000001e+24},{"model":"Qwen3-235B-A22B","date":"2025-04-28","score":139.35,"organization":"Alibaba","lab":null,"model_id":null,"flop":4.752e+24},{"model":"Qwen3-8B","date":"2025-04-28","score":136.18,"organization":"Alibaba","lab":null,"model_id":null,"flop":1.7712e+24},{"model":"Qwen3-32B","date":"2025-04-29","score":138.51,"organization":"Alibaba","lab":null,"model_id":null,"flop":7.0848e+24},{"model":"Qwen3-14B","date":"2025-04-29","score":138.24,"organization":"Alibaba","lab":null,"model_id":null,"flop":3.1968e+24},{"model":"Qwen3-30B-A3B","date":"2025-04-29","score":136.19,"organization":"Alibaba","lab":null,"model_id":null,"flop":6.48e+23},{"model":"DeepSeek-R1 (May 2025)","date":"2025-05-28","score":141.29,"organization":"DeepSeek","lab":null,"model_id":null,"flop":4.020010000000001e+24},{"model":"Kimi K2 (Jul 2025)","date":"2025-07-12","score":140.11,"organization":"Moonshot","lab":null,"model_id":null,"flop":2.976e+24},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","date":"2025-07-25","score":143.88,"organization":"Alibaba","lab":null,"model_id":null,"flop":4.752e+24},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","date":"2025-07-25","score":138.92,"organization":"Alibaba","lab":null,"model_id":null,"flop":4.752e+24},{"model":"Qwen3-30B-A3B-Instruct (Jul 2025)","date":"2025-07-29","score":137.42,"organization":"Alibaba","lab":null,"model_id":null,"flop":6.48e+23},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","date":"2025-07-30","score":139.64,"organization":"Alibaba","lab":null,"model_id":null,"flop":6.48e+23},{"model":"gpt-oss-120b","date":"2025-08-05","score":140.1,"organization":"OpenAI","lab":null,"model_id":null,"flop":4.94e+24},{"model":"gpt-oss-20b","date":"2025-08-05","score":137.8,"organization":"OpenAI","lab":null,"model_id":null,"flop":5.49e+23},{"model":"DeepSeek-V3.1","date":"2025-08-21","score":139.92,"organization":"DeepSeek","lab":null,"model_id":null,"flop":3.594058e+24},{"model":"DeepSeek-V3.2-Exp","date":"2025-09-29","score":145.02,"organization":"DeepSeek","lab":null,"model_id":null,"flop":4.18e+24},{"model":"GLM-4.6","date":"2025-09-30","score":140.78,"organization":"Z.ai (Zhipu AI),Tsinghua University","lab":null,"model_id":null,"flop":4.42e+24},{"model":"Kimi K2 Thinking","date":"2025-11-06","score":145.76,"organization":"Moonshot","lab":null,"model_id":null,"flop":4.2e+24},{"model":"DeepSeek-V3.2","date":"2025-12-01","score":146.19,"organization":"DeepSeek","lab":null,"model_id":null,"flop":4.2e+24},{"model":"GLM-4.7","date":"2025-12-22","score":143.46,"organization":"Z.ai (Zhipu AI)","lab":null,"model_id":null,"flop":4.42e+24},{"model":"Kimi K2.5","date":"2026-01-27","score":148.01,"organization":"Moonshot","lab":null,"model_id":null,"flop":5.8e+24},{"model":"GLM-5","date":"2026-02-11","score":145.85,"organization":"Z.ai (Zhipu AI)","lab":null,"model_id":null,"flop":6.84e+24},{"model":"DeepSeek-V4-Pro","date":"2026-04-24","score":149.09,"organization":"DeepSeek","lab":null,"model_id":null,"flop":9.702e+24},{"model":"DeepSeek-V4-Flash","date":"2026-04-24","score":146.11,"organization":"DeepSeek","lab":null,"model_id":null,"flop":2.496e+24},{"model":"Nemotron 3 Ultra","date":"2026-06-04","score":146.27,"organization":"NVIDIA","lab":null,"model_id":null,"flop":6.6e+24},{"model":"Inkling","date":"2026-07-15","score":148.64,"organization":"Thinking Machines","lab":null,"model_id":null,"flop":1.845e+24},{"model":"Kimi K3","date":"2026-07-16","score":157.63,"organization":"Moonshot","lab":null,"model_id":null,"flop":2.0001e+25},{"model":"DeepSeek V4 Flash 0731","date":"2026-07-31","score":154.49,"organization":"DeepSeek","lab":null,"model_id":null,"flop":2.496e+24},{"model":"DeepSeek V4 Pro 0813","date":"2026-08-13","score":155.47,"organization":"DeepSeek","lab":null,"model_id":null,"flop":9.702e+24},{"model":"GLM-5.3-Flash","date":"2026-08-20","score":151.45,"organization":"Z.ai (Zhipu AI)","lab":null,"model_id":null,"flop":3.24e+24}],"closed_share":{"open":126,"closed":140},"median_gap":5.4,"samples":12,"curve":[{"date":"2023-02-24","score":109.94,"model":"LLaMA-65B"},{"date":"2023-03-14","score":125.89,"model":"GPT-4 (Mar 2023)"},{"date":"2024-01-25","score":126.45,"model":"GPT-4 Turbo (Nov 2023)"},{"date":"2024-02-29","score":126.91,"model":"Claude 3 Opus"},{"date":"2024-04-09","score":127.25,"model":"GPT-4 Turbo (Apr 2024)"},{"date":"2024-05-13","score":128.97,"model":"GPT-4o (May 2024)"},{"date":"2024-06-20","score":130.0,"model":"Claude 3.5 Sonnet"},{"date":"2024-09-12","score":135.82,"model":"o1-mini"},{"date":"2024-12-17","score":141.86,"model":"o1"},{"date":"2025-03-31","score":144.16,"model":"Gemini 2.5 Pro (Mar 2025)"},{"date":"2025-04-16","score":146.91,"model":"o3"},{"date":"2025-06-10","score":147.45,"model":"o3-pro"},{"date":"2025-08-07","score":150.0,"model":"GPT-5"},{"date":"2025-10-07","score":150.29,"model":"GPT-5 Pro"},{"date":"2025-11-18","score":153.0,"model":"Gemini 3 Pro"},{"date":"2025-12-11","score":155.38,"model":"GPT-5.2 Pro"},{"date":"2026-02-05","score":156.58,"model":"GPT-5.3 Codex"},{"date":"2026-03-05","score":158.95,"model":"GPT-5.4 Pro"},{"date":"2026-04-23","score":162.25,"model":"GPT-5.5 Pro"},{"date":"2026-06-09","score":163.27,"model":"Claude Fable 5"},{"date":"2026-09-01","score":164.47,"model":"Claude Fable 5.1"},{"date":"2026-09-03","score":166.31,"model":"GPT-6 Astra"}],"tenure":[{"start":"2023-02-24","end":"2023-03-14","score":109.94,"models":["LLaMA-65B"]},{"start":"2023-03-14","end":"2024-01-25","score":125.89,"models":["GPT-4 (Mar 2023)"]},{"start":"2024-01-25","end":"2024-02-29","score":126.45,"models":["GPT-4 Turbo (Nov 2023)"]},{"start":"2024-02-29","end":"2024-04-09","score":126.91,"models":["Claude 3 Opus"]},{"start":"2024-04-09","end":"2024-05-13","score":127.25,"models":["GPT-4 Turbo (Apr 2024)"]},{"start":"2024-05-13","end":"2024-06-20","score":128.97,"models":["GPT-4o (May 2024)"]},{"start":"2024-06-20","end":"2024-09-12","score":130.0,"models":["Claude 3.5 Sonnet"]},{"start":"2024-09-12","end":"2024-12-17","score":135.82,"models":["o1-mini"]},{"start":"2024-12-17","end":"2025-03-31","score":141.86,"models":["o1"]},{"start":"2025-03-31","end":"2025-04-16","score":144.16,"models":["Gemini 2.5 Pro (Mar 2025)"]},{"start":"2025-04-16","end":"2025-06-10","score":146.91,"models":["o3"]},{"start":"2025-06-10","end":"2025-08-07","score":147.45,"models":["o3-pro"]},{"start":"2025-08-07","end":"2025-10-07","score":150.0,"models":["GPT-5"]},{"start":"2025-10-07","end":"2025-11-18","score":150.29,"models":["GPT-5 Pro"]},{"start":"2025-11-18","end":"2025-12-11","score":153.0,"models":["Gemini 3 Pro"]},{"start":"2025-12-11","end":"2026-02-05","score":155.38,"models":["GPT-5.2 Pro"]},{"start":"2026-02-05","end":"2026-03-05","score":156.58,"models":["GPT-5.3 Codex"]},{"start":"2026-03-05","end":"2026-04-23","score":158.95,"models":["GPT-5.4 Pro"]},{"start":"2026-04-23","end":"2026-06-09","score":162.25,"models":["GPT-5.5 Pro"]},{"start":"2026-06-09","end":"2026-09-01","score":163.27,"models":["Claude Fable 5"]},{"start":"2026-09-01","end":"2026-09-03","score":164.47,"models":["Claude Fable 5.1"]},{"start":"2026-09-03","end":"2026-09-15","score":166.31,"models":["GPT-6 Astra"]}],"as_of":"2026-09-15"},"efficiency":[{"id":"aider-polyglot","name":"Aider polyglot","source":"https://aider.chat/docs/leaderboards/","rows":[{"id":"aider_polyglot_external.csv:5","model":"GPT-5","version":"gpt-5-2025-08-07_low","effort":"low","score":81.3,"percent":null,"cost":null,"seconds":62.4,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"aider_polyglot_external.csv:6","model":"GPT-5","version":"gpt-5-2025-08-07_medium","effort":"medium","score":86.7,"percent":null,"cost":null,"seconds":118.7,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"aider_polyglot_external.csv:7","model":"GPT-5","version":"gpt-5-2025-08-07_high","effort":"high","score":88.0,"percent":null,"cost":null,"seconds":194.0,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"aider_polyglot_external.csv:9","model":"Grok-3 mini","version":"grok-3-mini-beta_low","effort":null,"score":34.7,"percent":null,"cost":null,"seconds":35.1,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:10","model":"openhands-lm-32b-v0.1","version":"openhands-lm-32b-v0.1","effort":null,"score":10.2,"percent":null,"cost":null,"seconds":195.6,"provider":null,"date":"2024-03-26","organization":null,"harness":"","revision":null},{"id":"aider_polyglot_external.csv:13","model":"o3-pro","version":"o3-pro-2025-06-10_high","effort":null,"score":84.9,"percent":null,"cost":null,"seconds":449.0,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"aider_polyglot_external.csv:14","model":"Gemini 2.5 Pro Preview (Jun 2025)","version":"gemini-2.5-pro-preview-05-06","effort":null,"score":76.9,"percent":null,"cost":null,"seconds":165.3,"provider":null,"date":"2025-05-06","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:15","model":"Gemini 2.5 Pro Preview (Jun 2025)","version":"gemini-2.5-pro-preview-06-05","effort":null,"score":79.1,"percent":null,"cost":null,"seconds":175.2,"provider":null,"date":"2025-06-05","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:16","model":"Gemini 2.5 Pro Preview (Jun 2025)","version":"gemini-2.5-pro-preview-06-05_32K","effort":null,"score":83.1,"percent":null,"cost":null,"seconds":200.3,"provider":null,"date":"2025-06-05","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:17","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_23K","effort":null,"score":55.1,"percent":null,"cost":null,"seconds":53.9,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:18","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20","effort":null,"score":44.0,"percent":null,"cost":null,"seconds":12.2,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:19","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514","effort":null,"score":56.4,"percent":null,"cost":null,"seconds":29.8,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"aider_polyglot_external.csv:20","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_32K","effort":null,"score":61.3,"percent":null,"cost":null,"seconds":79.9,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"aider_polyglot_external.csv:21","model":"Claude Opus 4","version":"claude-opus-4-20250514","effort":null,"score":70.7,"percent":null,"cost":null,"seconds":42.5,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"aider_polyglot_external.csv:22","model":"Claude Opus 4","version":"claude-opus-4-20250514_32K","effort":null,"score":72.0,"percent":null,"cost":null,"seconds":44.1,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"aider_polyglot_external.csv:23","model":"Qwen3-32B","version":"Qwen3-32B","effort":null,"score":40.0,"percent":null,"cost":null,"seconds":372.2,"provider":null,"date":"2025-04-29","organization":"Alibaba","harness":"","revision":null},{"id":"aider_polyglot_external.csv:24","model":"Llama 4 Maverick","version":"Llama-4-Maverick-17B-128E-Instruct","effort":null,"score":15.6,"percent":null,"cost":null,"seconds":20.5,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:25","model":"GPT-4.1 mini","version":"gpt-4.1-mini-2025-04-14","effort":null,"score":32.4,"percent":null,"cost":null,"seconds":19.5,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"id":"aider_polyglot_external.csv:26","model":"GPT-4.1 nano","version":"gpt-4.1-nano-2025-04-14","effort":null,"score":8.9,"percent":null,"cost":null,"seconds":12.0,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"id":"aider_polyglot_external.csv:27","model":"GPT-4o (Mar 2025)","version":"chatgpt-4o-03-27-2025","effort":null,"score":45.3,"percent":null,"cost":null,"seconds":10.3,"provider":null,"date":"2025-03-27","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:28","model":"Gemini 2.5 Flash Preview (Apr 2025)","version":"gemini-2.5-flash-preview-04-17","effort":null,"score":47.1,"percent":null,"cost":null,"seconds":50.1,"provider":null,"date":"2025-04-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:29","model":"Grok-3 mini","version":"grok-3-mini-beta_high","effort":null,"score":49.3,"percent":null,"cost":null,"seconds":79.1,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:30","model":"GPT-4.1","version":"gpt-4.1-2025-04-14","effort":null,"score":52.4,"percent":null,"cost":null,"seconds":20.5,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"id":"aider_polyglot_external.csv:31","model":"Grok 3","version":"grok-3-beta","effort":null,"score":53.3,"percent":null,"cost":null,"seconds":15.3,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:32","model":"DeepSeek-V3 (Mar 2025)","version":"DeepSeek-V3-0324","effort":null,"score":55.1,"percent":null,"cost":null,"seconds":290.0,"provider":null,"date":"2025-03-24","organization":"DeepSeek","harness":"","revision":null},{"id":"aider_polyglot_external.csv:33","model":"DeepSeek-R1 (May 2025)","version":"DeepSeek-R1-0528","effort":null,"score":71.4,"percent":null,"cost":null,"seconds":716.6,"provider":null,"date":"2025-05-28","organization":"DeepSeek","harness":"","revision":null},{"id":"aider_polyglot_external.csv:35","model":"o4-mini","version":"o4-mini-2025-04-16_high","effort":"high","score":72.0,"percent":null,"cost":null,"seconds":176.5,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"aider_polyglot_external.csv:36","model":"o3","version":"o3-2025-04-16_high","effort":"high","score":81.3,"percent":null,"cost":null,"seconds":197.3,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"aider_polyglot_external.csv:38","model":"Gemini 2.0 Flash Thinking Exp","version":"gemini-2.0-flash-thinking-exp-01-21","effort":null,"score":18.2,"percent":null,"cost":null,"seconds":24.2,"provider":null,"date":"2025-01-21","organization":"Google DeepMind,Google","harness":"","revision":null},{"id":"aider_polyglot_external.csv:39","model":"GPT-4o (Nov 2024)","version":"gpt-4o-2024-11-20","effort":null,"score":18.2,"percent":null,"cost":null,"seconds":12.1,"provider":null,"date":"2024-11-20","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:40","model":"Claude 3.5 Sonnet (Oct 2024)","version":"claude-3-5-sonnet-20241022","effort":null,"score":51.6,"percent":null,"cost":null,"seconds":21.4,"provider":null,"date":"2024-10-22","organization":"Anthropic","harness":"","revision":null},{"id":"aider_polyglot_external.csv:42","model":"Claude 3.5 Haiku (Oct 2024)","version":"claude-3-5-haiku-20241022","effort":null,"score":28.0,"percent":null,"cost":null,"seconds":31.8,"provider":null,"date":"2024-10-22","organization":"Anthropic","harness":"","revision":null},{"id":"aider_polyglot_external.csv:43","model":"Gemini 2.0 Flash (Dec 2024)","version":"gemini-exp-1206","effort":null,"score":38.2,"percent":null,"cost":null,"seconds":45.5,"provider":null,"date":"2024-12-06","organization":"Google DeepMind,Google","harness":"","revision":null},{"id":"aider_polyglot_external.csv:44","model":"DeepSeek-R1","version":"DeepSeek-R1","effort":null,"score":56.9,"percent":null,"cost":null,"seconds":113.7,"provider":null,"date":"2025-01-20","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"id":"aider_polyglot_external.csv:46","model":"Qwen2.5-Max","version":"qwen-max-2025-01-25","effort":null,"score":21.8,"percent":null,"cost":null,"seconds":39.5,"provider":null,"date":"2025-01-25","organization":"Alibaba","harness":"","revision":null},{"id":"aider_polyglot_external.csv:47","model":"Gemini 2.0 Pro Exp (Feb 2025)","version":"gemini-2.0-pro-exp-02-05","effort":null,"score":35.6,"percent":null,"cost":null,"seconds":34.8,"provider":null,"date":"2025-02-05","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:49","model":"QwQ-32B","version":"QwQ-32B","effort":null,"score":20.9,"percent":null,"cost":null,"seconds":228.6,"provider":null,"date":"2025-03-05","organization":"Alibaba","harness":"","revision":null,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"id":"aider_polyglot_external.csv:50","model":"Cohere Command A","version":"c4ai-command-a-03-2025","effort":null,"score":12.0,"percent":null,"cost":null,"seconds":85.1,"provider":null,"date":"2025-03-13","organization":"Cohere","harness":"","revision":null},{"id":"aider_polyglot_external.csv:51","model":"GPT-4o mini","version":"gpt-4o-mini-2024-07-18","effort":null,"score":3.6,"percent":null,"cost":null,"seconds":17.3,"provider":null,"date":"2024-07-18","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"id":"aider_polyglot_external.csv:52","model":"Gemma 3 27B","version":"gemma-3-27b-it","effort":null,"score":4.9,"percent":null,"cost":null,"seconds":79.7,"provider":null,"date":"2025-03-12","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:53","model":"GPT-4.5 Preview (Feb 2025)","version":"gpt-4.5-preview-2025-02-27","effort":null,"score":44.9,"percent":null,"cost":null,"seconds":113.5,"provider":null,"date":"2025-02-27","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:55","model":"Claude 3.7 Sonnet","version":"claude-3-7-sonnet-20250219","effort":null,"score":60.4,"percent":null,"cost":null,"seconds":28.3,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"aider_polyglot_external.csv:56","model":"Gemini 2.0 Flash (Dec 2024)","version":"gemini-2.0-flash-exp","effort":null,"score":22.2,"percent":null,"cost":null,"seconds":12.2,"provider":null,"date":"2024-12-11","organization":"Google DeepMind,Google","harness":"","revision":null},{"id":"aider_polyglot_external.csv:57","model":"DeepSeek-V2.5 (Sep 2024)","version":"DeepSeek-V2.5","effort":null,"score":17.8,"percent":null,"cost":null,"seconds":184.0,"provider":null,"date":"2024-09-06","organization":"DeepSeek","harness":"","revision":null},{"id":"aider_polyglot_external.csv:59","model":"GPT-4o (Jan 2025)","version":"chatgpt-4o-01-29-2025","effort":null,"score":27.1,"percent":null,"cost":null,"seconds":12.4,"provider":null,"date":"2025-01-29","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:60","model":"Codestral","version":"codestral-2501","effort":null,"score":11.1,"percent":null,"cost":null,"seconds":9.3,"provider":null,"date":"2025-01-13","organization":"Mistral AI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:61","model":"o1","version":"o1-2024-12-17_high","effort":"high","score":61.7,"percent":null,"cost":null,"seconds":133.2,"provider":null,"date":"2024-12-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"id":"aider_polyglot_external.csv:62","model":"o3-mini","version":"o3-mini-2025-01-31_medium","effort":"medium","score":53.8,"percent":null,"cost":null,"seconds":47.2,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"aider_polyglot_external.csv:63","model":"Yi-Lightning","version":"yi-lightning","effort":null,"score":12.9,"percent":null,"cost":null,"seconds":146.7,"provider":null,"date":"2024-12-02","organization":"01.AI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:64","model":"GPT-4o (Aug 2024)","version":"gpt-4o-2024-08-06","effort":null,"score":23.1,"percent":null,"cost":null,"seconds":16.0,"provider":null,"date":"2024-08-06","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:66","model":"Claude 3.7 Sonnet (32k thinking)","version":"claude-3-7-sonnet-20250219_32K","effort":null,"score":64.9,"percent":null,"cost":null,"seconds":105.2,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"aider_polyglot_external.csv:67","model":"o3-mini","version":"o3-mini-2025-01-31_high","effort":"high","score":60.4,"percent":null,"cost":null,"seconds":124.6,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"aider_polyglot_external.csv:68","model":"o1-mini","version":"o1-mini-2024-09-12_unknown","effort":"unknown thinking","score":32.9,"percent":null,"cost":null,"seconds":34.7,"provider":null,"date":"2024-09-12","organization":"OpenAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:69","model":"Gemini 2.5 Pro Preview (Mar 2025)","version":"gemini-2.5-pro-preview-03-25","effort":null,"score":72.9,"percent":null,"cost":null,"seconds":45.3,"provider":null,"date":"2025-03-31","organization":"Google DeepMind","harness":"","revision":null},{"id":"aider_polyglot_external.csv:70","model":"Qwen3-235B-A22B-Instruct (Jul 2025)","version":"Qwen3-235B-A22B-Instruct-2507","effort":null,"score":59.6,"percent":null,"cost":null,"seconds":45.4,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"aider_polyglot_external.csv:71","model":"o3","version":"o3-2025-04-16_unknown","effort":"unknown thinking","score":76.9,"percent":null,"cost":null,"seconds":101.7,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"aider_polyglot_external.csv:72","model":"Grok 4","version":"grok-4-0709_high","effort":null,"score":79.6,"percent":null,"cost":null,"seconds":403.2,"provider":null,"date":"2025-07-09","organization":"xAI","harness":"","revision":null},{"id":"aider_polyglot_external.csv:73","model":"Kimi K2 0905 (Novita)","version":"moonshotai/kimi-k2-0905","effort":null,"score":59.1,"percent":null,"cost":null,"seconds":67.6,"provider":null,"date":"2025-09-05","organization":"Moonshot","harness":"","revision":null},{"id":"aider_polyglot_external.csv:74","model":"gpt-oss-120b","version":"openai/gpt-oss-120b_high","effort":null,"score":41.8,"percent":null,"cost":null,"seconds":35.5,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"id":"aider_polyglot_external.csv:75","model":"DeepSeek-V3.2","version":"deepseek-reasoner","effort":"thinking","score":74.2,"percent":null,"cost":null,"seconds":291.2,"provider":null,"date":"2025-12-01","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"id":"aider_polyglot_external.csv:76","model":"DeepSeek-V3.2","version":"deepseek-chat","effort":null,"score":70.2,"percent":null,"cost":null,"seconds":104.0,"provider":null,"date":"2025-12-01","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"}],"cost_label":null,"time_label":"Time per case (seconds)"},{"id":"arc-agi","name":"ARC-AGI","source":"https://arcprize.org/leaderboard","rows":[{"id":"arc_agi_external.csv:4","model":"Gemini 3.1 Pro Preview","version":"gemini-3.1-pro-preview","effort":null,"score":0.98,"percent":98.0,"cost":0.5223,"seconds":null,"provider":null,"date":"2026-02-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"id":"arc_agi_external.csv:5","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_xhigh","effort":"xhigh","score":0.975,"percent":97.5,"cost":0.4,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_external.csv:13","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_high","effort":"high","score":0.97,"percent":97.0,"cost":0.3,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_external.csv:14","model":"GPT-5.5 Pro","version":"gpt-5.5-pro_high","effort":"high","score":0.965,"percent":96.5,"cost":4.53,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"id":"arc_agi_external.csv:15","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_max","effort":"max","score":0.965,"percent":96.5,"cost":0.54,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_external.csv:16","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_max","effort":"max","score":0.965,"percent":96.5,"cost":0.55,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_external.csv:24","model":"GPT-5.5","version":"gpt-5.5_xhigh","effort":"xhigh","score":0.9500000000000001,"percent":95.0,"cost":0.73,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_external.csv:25","model":"GPT-5.5 Pro","version":"gpt-5.5-pro_xhigh","effort":"xhigh","score":0.9500000000000001,"percent":95.0,"cost":4.52,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"id":"arc_agi_external.csv:28","model":"GPT-5.4 Pro","version":"gpt-5.4-pro-2026-03-05_xhigh","effort":"xhigh","score":0.9450000000000001,"percent":94.5,"cost":7.21,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"id":"arc_agi_external.csv:29","model":"GPT-5.5","version":"gpt-5.5_high","effort":"high","score":0.9450000000000001,"percent":94.5,"cost":0.56,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_external.csv:30","model":"Kimi K3","version":"kimi-k3_max","effort":"max","score":0.9450000000000001,"percent":94.5,"cost":0.7704815999999995,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_external.csv:33","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_xhigh","effort":"xhigh","score":0.9400000000000001,"percent":94.0,"cost":0.26,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_external.csv:37","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_xhigh","effort":"xhigh","score":0.9367000000000001,"percent":93.67000000000002,"cost":0.62,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_external.csv:39","model":"Claude Opus 4.7","version":"claude-opus-4-7_high","effort":"high","score":0.935,"percent":93.5,"cost":1.41,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_external.csv:42","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_high","effort":"high","score":0.9267000000000001,"percent":92.67,"cost":0.37,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_external.csv:43","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_high","effort":null,"score":0.925,"percent":92.5,"cost":0.4278,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"arc_agi_external.csv:44","model":"Claude Opus 4.8","version":"claude-opus-4-8_max","effort":null,"score":0.925,"percent":92.5,"cost":2.332,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_external.csv:45","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_medium","effort":"medium","score":0.925,"percent":92.5,"cost":0.22,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_external.csv:47","model":"GPT-5.5","version":"gpt-5.5_medium","effort":"medium","score":0.9217000000000001,"percent":92.17,"cost":0.39,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_external.csv:48","model":"Claude Opus 4.7","version":"claude-opus-4-7_max","effort":"max","score":0.92,"percent":92.0,"cost":2.58,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_external.csv:49","model":"Claude Opus 4.8","version":"claude-opus-4-8_high","effort":null,"score":0.92,"percent":92.0,"cost":1.0416,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_external.csv:50","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_high","effort":"high","score":0.92,"percent":92.0,"cost":0.19,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_external.csv:53","model":"Claude Opus 4.8","version":"claude-opus-4-8_medium","effort":null,"score":0.915,"percent":91.5,"cost":0.9115,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_external.csv:56","model":"Claude Opus 4.7","version":"claude-opus-4-7_low","effort":"low","score":0.91,"percent":91.0,"cost":0.76,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_external.csv:57","model":"Claude Opus 4.7","version":"claude-opus-4-7_medium","effort":"medium","score":0.91,"percent":91.0,"cost":1.04,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_external.csv:58","model":"GPT-5.2 Pro","version":"gpt-5.2-pro-2025-12-11_xhigh","effort":null,"score":0.905,"percent":90.5,"cost":11.6542,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"id":"arc_agi_external.csv:65","model":"Grok 4.20","version":"grok-4-20","effort":null,"score":0.895,"percent":89.5,"cost":0.35,"seconds":null,"provider":null,"date":"2026-02-17","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_external.csv:67","model":"Claude Opus 4.8","version":"claude-opus-4-8_low","effort":null,"score":0.88,"percent":88.0,"cost":0.6711,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_external.csv:68","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_max","effort":"max","score":0.88,"percent":88.0,"cost":0.32,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_external.csv:69","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_xhigh","effort":"xhigh","score":0.8767,"percent":87.67,"cost":0.21,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_external.csv:71","model":"Gemini 3 Deep Think","version":"gemini-3-deep-think-preview","effort":null,"score":0.875,"percent":87.5,"cost":77.16,"seconds":null,"provider":null,"date":"2026-02-12","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:74","model":"Grok 4.5","version":"grok-4.5_medium","effort":"medium","score":0.8717,"percent":87.17,"cost":0.31,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_external.csv:79","model":"Kimi K3","version":"kimi-k3_high","effort":"high","score":0.8666666666666666,"percent":86.66666666666666,"cost":0.47977101000000005,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_external.csv:80","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_high","effort":"high","score":0.865,"percent":86.5,"cost":1.1908,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"arc_agi_external.csv:81","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_xhigh","effort":"xhigh","score":0.862,"percent":86.2,"cost":0.96,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_external.csv:82","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_medium","effort":"medium","score":0.8617,"percent":86.17,"cost":0.25,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_external.csv:85","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_max","effort":"max","score":0.86,"percent":86.0,"cost":1.452,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"arc_agi_external.csv:88","model":"GPT-5.2 Pro","version":"gpt-5.2-pro-2025-12-11_high","effort":null,"score":0.8567,"percent":85.67,"cost":5.8694,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"id":"arc_agi_external.csv:89","model":"Grok 4.5","version":"grok-4.5_high","effort":"high","score":0.8567,"percent":85.67,"cost":0.33,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_external.csv:91","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":null,"score":0.8467,"percent":84.67,"cost":0.1743,"seconds":null,"provider":null,"date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"arc_agi_external.csv:95","model":"GPT-5.2 Pro","version":"gpt-5.2-pro-2025-12-11_medium","effort":null,"score":0.8117000000000001,"percent":81.17000000000002,"cost":3.9774,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"id":"arc_agi_external.csv:96","model":"Claude Opus 4.5 (64k thinking)","version":"claude-opus-4-5-20251101_64K","effort":null,"score":0.8,"percent":80.0,"cost":1.47,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:98","model":"Inkling","version":"Inkling","effort":null,"score":0.795,"percent":79.5,"cost":0.30047416199999993,"seconds":null,"provider":null,"date":"2026-07-15","organization":"Thinking Machines","harness":"","revision":null},{"id":"arc_agi_external.csv:99","model":"Grok 4.5","version":"grok-4.5_low","effort":"low","score":0.7917000000000001,"percent":79.17,"cost":0.18,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_external.csv:100","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_high","effort":"high","score":0.7867000000000001,"percent":78.67,"cost":0.5196,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_external.csv:102","model":"GLM-5.2","version":"glm-5.2_unknown","effort":null,"score":0.77,"percent":77.0,"cost":0.1916,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"arc_agi_external.csv:103","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_medium","effort":"medium","score":0.77,"percent":77.0,"cost":0.13,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_external.csv:104","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_high","effort":"high","score":0.765,"percent":76.5,"cost":0.13,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_external.csv:106","model":"GPT-5.5","version":"gpt-5.5_low","effort":"low","score":0.7617,"percent":76.17,"cost":0.2,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_external.csv:108","model":"Claude Opus 4.5 (32k thinking)","version":"claude-opus-4-5-20251101_32K","effort":null,"score":0.758,"percent":75.8,"cost":0.95,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:110","model":"Gemini 3 Pro Preview","version":"gemini-3-pro-preview","effort":null,"score":0.75,"percent":75.0,"cost":0.4926,"seconds":null,"provider":null,"date":"2025-11-18","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"id":"arc_agi_external.csv:112","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_low","effort":"low","score":0.745,"percent":74.5,"cost":0.17,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_external.csv:114","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_high","effort":"high","score":0.728,"percent":72.8,"cost":0.674,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_external.csv:115","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_medium","effort":"medium","score":0.7267,"percent":72.67,"cost":0.3448,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_external.csv:116","model":"Claude Opus 4.5 (16k thinking)","version":"claude-opus-4-5-20251101_16K","effort":null,"score":0.72,"percent":72.0,"cost":0.62,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:118","model":"GPT-5 Pro","version":"gpt-5-pro-2025-10-06_unknown","effort":null,"score":0.7017,"percent":70.17,"cost":4.7785,"seconds":null,"provider":null,"date":"2025-10-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"id":"arc_agi_external.csv:119","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_low","effort":"low","score":0.6817000000000001,"percent":68.17,"cost":0.15,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_external.csv:121","model":"Grok 4","version":"grok-4-0709","effort":null,"score":0.6667000000000001,"percent":66.67,"cost":2.17,"seconds":null,"provider":null,"date":"2025-07-09","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_external.csv:122","model":"GPT-5","version":"gpt-5-2025-08-07_high","effort":"high","score":0.6567000000000001,"percent":65.67,"cost":0.5087,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_external.csv:123","model":"Kimi K3","version":"kimi-k3_low","effort":"low","score":0.6566666666666666,"percent":65.66666666666666,"cost":0.18061866,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_external.csv:124","model":"Kimi K2.5","version":"kimi-k2.5","effort":null,"score":0.6533,"percent":65.33,"cost":0.15,"seconds":null,"provider":null,"date":"2026-01-27","organization":"Moonshot","harness":"","revision":null},{"id":"arc_agi_external.csv:125","model":"Claude Sonnet 4.5 (32k thinking)","version":"claude-sonnet-4-5-20250929_32K","effort":null,"score":0.6367,"percent":63.67,"cost":0.5193,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:126","model":"MiniMax-M2.5","version":"MiniMax-M2.5","effort":null,"score":0.6367,"percent":63.67,"cost":0.07,"seconds":null,"provider":null,"date":"2026-02-12","organization":"MiniMax","harness":"","revision":null},{"id":"arc_agi_external.csv:127","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_xhigh","effort":"xhigh","score":0.6367,"percent":63.67,"cost":0.51,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_external.csv:129","model":"o3","version":"o3-2025-04-16_high","effort":"high","score":0.6083,"percent":60.83,"cost":0.5002,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_external.csv:130","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_low","effort":"low","score":0.6017,"percent":60.17,"cost":0.09,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_external.csv:131","model":"o3-pro","version":"o3-pro-2025-06-10_high","effort":null,"score":0.5933,"percent":59.330000000000005,"cost":4.16,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_external.csv:132","model":"o4-mini","version":"o4-mini-2025-04-16_high","effort":"high","score":0.5867,"percent":58.67,"cost":0.4058,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_external.csv:133","model":"Claude Opus 4.5 (8k thinking)","version":"claude-opus-4-5-20251101_8K","effort":null,"score":0.5867,"percent":58.67,"cost":0.38,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:134","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_high","effort":"high","score":0.58,"percent":57.99999999999999,"cost":0.27,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_external.csv:135","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_medium","effort":"medium","score":0.577,"percent":57.699999999999996,"cost":0.3149,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_external.csv:136","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":null,"score":0.5767,"percent":57.67,"cost":0.0612,"seconds":null,"provider":null,"date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"arc_agi_external.csv:138","model":"o3-pro","version":"o3-pro-2025-06-10_medium","effort":null,"score":0.57,"percent":56.99999999999999,"cost":3.1766,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_external.csv:139","model":"DeepSeek-V3.2 (Thinking; Novita)","version":"deepseek/deepseek-v3.2","effort":null,"score":0.57,"percent":56.99999999999999,"cost":0.08,"seconds":null,"provider":null,"date":"2025-12-01","organization":"DeepSeek","harness":"","revision":null},{"id":"arc_agi_external.csv:140","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_medium","effort":"medium","score":0.565,"percent":56.49999999999999,"cost":0.06,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_external.csv:141","model":"GPT-5","version":"gpt-5-2025-08-07_medium","effort":"medium","score":0.5617,"percent":56.169999999999995,"cost":0.3301,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_external.csv:142","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_low","effort":"low","score":0.5567,"percent":55.669999999999995,"cost":0.1657,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_external.csv:143","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_high","effort":"high","score":0.5433,"percent":54.33,"cost":0.116,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_external.csv:144","model":"o3","version":"o3-2025-04-16_medium","effort":"medium","score":0.5383,"percent":53.83,"cost":0.2882,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_external.csv:147","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_xhigh","effort":"xhigh","score":0.515,"percent":51.5,"cost":0.13,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_external.csv:149","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_minimal","effort":null,"score":0.4883,"percent":48.83,"cost":0.0645,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"arc_agi_external.csv:150","model":"Grok 4 Fast","version":"grok-4-fast","effort":null,"score":0.485,"percent":48.5,"cost":0.0307,"seconds":null,"provider":null,"date":"2025-09-19","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_external.csv:152","model":"Claude Sonnet 4.5 (16k thinking)","version":"claude-sonnet-4-5-20250929_16K","effort":null,"score":0.4833,"percent":48.33,"cost":0.2683,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:153","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_32K","effort":null,"score":0.4767,"percent":47.67,"cost":0.2558,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_external.csv:154","model":"Claude Sonnet 4.5 (8k thinking)","version":"claude-sonnet-4-5-20250929_8K","effort":null,"score":0.465,"percent":46.5,"cost":0.1776,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:155","model":"GLM-5","version":"glm-5","effort":null,"score":0.44670000000000004,"percent":44.67,"cost":0.17,"seconds":null,"provider":null,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"id":"arc_agi_external.csv:156","model":"o3-pro","version":"o3-pro-2025-06-10_low","effort":null,"score":0.4433000000000001,"percent":44.330000000000005,"cost":1.6382,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_external.csv:157","model":"GPT-5","version":"gpt-5-2025-08-07_low","effort":"low","score":0.44,"percent":44.0,"cost":0.1531,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_external.csv:158","model":"o4-mini","version":"o4-mini-2025-04-16_medium","effort":"medium","score":0.4183,"percent":41.83,"cost":0.15,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_external.csv:159","model":"o3","version":"o3-2025-04-16_low","effort":"low","score":0.415,"percent":41.5,"cost":0.1764,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_external.csv:160","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_16K","effort":null,"score":0.41000000000000003,"percent":41.0,"cost":0.4839,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:161","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_medium","effort":"medium","score":0.4083,"percent":40.83,"cost":0.16,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_external.csv:162","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_16K","effort":null,"score":0.4,"percent":40.0,"cost":0.3658,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_external.csv:163","model":"tiny-recursion-model","version":"tiny-recursion-model","effort":null,"score":0.4,"percent":40.0,"cost":1.7638,"seconds":null,"provider":null,"date":"2025-10-06","organization":null,"harness":"","revision":null},{"id":"arc_agi_external.csv:164","model":"Claude Opus 4.5 (no thinking)","version":"claude-opus-4-5-20251101","effort":null,"score":0.4,"percent":40.0,"cost":0.13,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:165","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_high","effort":"high","score":0.38170000000000004,"percent":38.17,"cost":0.07,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_external.csv:166","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_medium","effort":"medium","score":0.3733000000000001,"percent":37.330000000000005,"cost":0.0401,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_external.csv:167","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_16K","effort":null,"score":0.3733000000000001,"percent":37.330000000000005,"cost":0.1042,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_external.csv:168","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_32K","effort":null,"score":0.37,"percent":37.0,"cost":0.5123,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:169","model":"Claude Opus 4","version":"claude-opus-4-20250514_16K","effort":null,"score":0.3567,"percent":35.67,"cost":1.2496,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_external.csv:171","model":"o3-mini","version":"o3-mini-2025-01-31_high","effort":"high","score":0.345,"percent":34.5,"cost":0.3989,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_external.csv:172","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_low","effort":"low","score":0.3417,"percent":34.17,"cost":0.04,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_external.csv:173","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_16K","effort":null,"score":0.3333,"percent":33.33,"cost":0.2134,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:174","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20","effort":null,"score":0.3333,"percent":33.33,"cost":0.0371,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:175","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_low","effort":"low","score":0.33170000000000005,"percent":33.17,"cost":0.1,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_external.csv:178","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_medium","effort":"medium","score":0.33,"percent":33.0,"cost":0.04,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_external.csv:180","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_23K","effort":null,"score":0.3233,"percent":32.33,"cost":0.1971,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:182","model":"Gemini 2.5 Pro Preview (Jun 2025)","version":"gemini-2.5-pro-preview-06-05_1K","effort":null,"score":0.313,"percent":31.3,"cost":0.0573,"seconds":null,"provider":null,"date":"2025-06-05","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:183","model":"Claude Sonnet 4.5 (1k thinking)","version":"claude-sonnet-4-5-20250929_1K","effort":null,"score":0.31,"percent":31.0,"cost":0.0928,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:185","model":"Claude Opus 4","version":"claude-opus-4-20250514_8K","effort":null,"score":0.3067,"percent":30.669999999999998,"cost":0.7408,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_external.csv:186","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_8K","effort":null,"score":0.295,"percent":29.5,"cost":0.2947,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:187","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_8K","effort":null,"score":0.29,"percent":28.999999999999996,"cost":0.1952,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_external.csv:188","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":null,"score":0.29,"percent":28.999999999999996,"cost":0.0163,"seconds":null,"provider":null,"date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"arc_agi_external.csv:189","model":"Claude 3.7 Sonnet (16k thinking)","version":"claude-3-7-sonnet-20250219_16K","effort":null,"score":0.286,"percent":28.599999999999998,"cost":0.33,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:190","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_1K","effort":null,"score":0.28,"percent":28.000000000000004,"cost":0.0937,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_external.csv:191","model":"codex-mini-2025-05-16","version":"codex-mini-2025-05-16","effort":null,"score":0.2733,"percent":27.33,"cost":0.1597,"seconds":null,"provider":null,"date":"2025-05-16","organization":null,"harness":"","revision":null},{"id":"arc_agi_external.csv:193","model":"Claude Opus 4","version":"claude-opus-4-20250514_1K","effort":null,"score":0.27,"percent":27.0,"cost":0.5021,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_external.csv:194","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_low","effort":"low","score":0.2633,"percent":26.33,"cost":0.0135,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_external.csv:195","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_8K","effort":null,"score":0.2583,"percent":25.83,"cost":0.1344,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:196","model":"Claude Sonnet 4.5 (no thinking)","version":"claude-sonnet-4-5-20250929","effort":null,"score":0.255,"percent":25.5,"cost":0.0808,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:197","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_8K","effort":null,"score":0.255,"percent":25.5,"cost":0.0687,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_external.csv:198","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514","effort":null,"score":0.2383,"percent":23.830000000000002,"cost":0.0806,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_external.csv:200","model":"Claude Opus 4","version":"claude-opus-4-20250514","effort":null,"score":0.225,"percent":22.5,"cost":0.4036,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_external.csv:201","model":"o3-mini","version":"o3-mini-2025-01-31_medium","effort":"medium","score":0.2233,"percent":22.33,"cost":0.1907,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_external.csv:202","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":null,"score":0.215,"percent":21.5,"cost":0.012,"seconds":null,"provider":null,"date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"arc_agi_external.csv:203","model":"o4-mini","version":"o4-mini-2025-04-16_low","effort":"low","score":0.2133,"percent":21.33,"cost":0.0406,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_external.csv:204","model":"DeepSeek-R1 (May 2025)","version":"DeepSeek-R1-0528","effort":null,"score":0.2121,"percent":21.21,"cost":0.0464,"seconds":null,"provider":null,"date":"2025-05-28","organization":"DeepSeek","harness":"","revision":null},{"id":"arc_agi_external.csv:205","model":"Claude 3.7 Sonnet (8k thinking)","version":"claude-3-7-sonnet-20250219_8K","effort":null,"score":0.212,"percent":21.2,"cost":0.21,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:206","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_medium","effort":"medium","score":0.2071,"percent":20.71,"cost":0.0124,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_external.csv:207","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_low","effort":"low","score":0.1833,"percent":18.33,"cost":0.01,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_external.csv:210","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_1K","effort":null,"score":0.16830000000000003,"percent":16.830000000000002,"cost":0.0306,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_external.csv:211","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_high","effort":"high","score":0.1667,"percent":16.669999999999998,"cost":0.0292,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_external.csv:213","model":"Grok 3 mini","version":"grok-3-mini_low","effort":null,"score":0.165,"percent":16.5,"cost":0.0099,"seconds":null,"provider":null,"date":"2025-06-24","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_external.csv:214","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_1K","effort":null,"score":0.16,"percent":16.0,"cost":0.0356,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_external.csv:216","model":"DeepSeek-R1","version":"DeepSeek-R1","effort":null,"score":0.158,"percent":15.8,"cost":0.06,"seconds":null,"provider":null,"date":"2025-01-20","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"id":"arc_agi_external.csv:217","model":"o3-mini","version":"o3-mini-2025-01-31_low","effort":null,"score":0.145,"percent":14.499999999999998,"cost":0.0519,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_external.csv:218","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001","effort":null,"score":0.1433,"percent":14.330000000000002,"cost":0.0263,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_external.csv:220","model":"o1-mini","version":"o1-mini-2024-09-12_unknown","effort":"unknown thinking","score":0.14,"percent":14.000000000000002,"cost":0.135,"seconds":null,"provider":null,"date":"2024-09-12","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_external.csv:221","model":"Claude 3.7 Sonnet","version":"claude-3-7-sonnet-20250219","effort":null,"score":0.136,"percent":13.600000000000001,"cost":0.058,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:222","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_low","effort":"low","score":0.13,"percent":13.0,"cost":0.04,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_external.csv:224","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_none","effort":"none","score":0.1233,"percent":12.33,"cost":0.082,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_external.csv:227","model":"Claude 3.7 Sonnet (1k thinking)","version":"claude-3-7-sonnet-20250219_1K","effort":null,"score":0.116,"percent":11.600000000000001,"cost":0.07,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_external.csv:228","model":"Qwen3-235B-A22B-Instruct (Jul 2025)","version":"Qwen3-235B-A22B-Instruct-2507","effort":null,"score":0.11,"percent":11.0,"cost":0.0025,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"arc_agi_external.csv:229","model":"GPT-4.5 Preview (Feb 2025)","version":"gpt-4.5-preview-2025-02-27","effort":null,"score":0.103,"percent":10.299999999999999,"cost":0.29,"seconds":null,"provider":null,"date":"2025-02-27","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_external.csv:231","model":"Magistral Medium 1.1","version":"magistral-medium-2506","effort":null,"score":0.061200000000000004,"percent":6.12,"cost":0.123,"seconds":null,"provider":null,"date":"2025-06-10","organization":"Mistral AI","harness":"","revision":null},{"id":"arc_agi_external.csv:232","model":"GPT-5","version":"gpt-5-2025-08-07_minimal","effort":"minimal","score":0.06,"percent":6.0,"cost":0.0335,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_external.csv:233","model":"Magistral Medium 1.1","version":"magistral-medium-2506","effort":null,"score":0.0591,"percent":5.91,"cost":0.108,"seconds":null,"provider":null,"date":"2025-06-10","organization":"Mistral AI","harness":"","revision":null},{"id":"arc_agi_external.csv:234","model":"GPT-5.1 (no thinking)","version":"gpt-5.1-2025-11-13_none","effort":null,"score":0.058300000000000005,"percent":5.83,"cost":0.0338,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_external.csv:235","model":"GPT-4.1","version":"gpt-4.1-2025-04-14","effort":null,"score":0.055,"percent":5.5,"cost":0.039,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"id":"arc_agi_external.csv:236","model":"Grok 3","version":"grok-3","effort":null,"score":0.055,"percent":5.5,"cost":0.0931,"seconds":null,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_external.csv:238","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_minimal","effort":"minimal","score":0.0533,"percent":5.33,"cost":0.0057,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_external.csv:239","model":"Magistral Small 1.0","version":"magistral-small-2506","effort":null,"score":0.05,"percent":5.0,"cost":0.0399,"seconds":null,"provider":null,"date":"2025-06-10","organization":"Mistral AI","harness":"","revision":null},{"id":"arc_agi_external.csv:240","model":"GPT-4o (Nov 2024)","version":"gpt-4o-2024-11-20","effort":null,"score":0.045,"percent":4.5,"cost":0.05,"seconds":null,"provider":null,"date":"2024-11-20","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_external.csv:241","model":"Llama 4 Maverick","version":"Llama-4-Maverick-17B-128E-Instruct","effort":null,"score":0.0438,"percent":4.38,"cost":0.0078,"seconds":null,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"arc_agi_external.csv:242","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_low","effort":"low","score":0.0404,"percent":4.04,"cost":0.0033,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_external.csv:243","model":"GPT-4.1 mini","version":"gpt-4.1-mini-2025-04-14","effort":null,"score":0.035,"percent":3.5000000000000004,"cost":0.0078,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"id":"arc_agi_external.csv:244","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_minimal","effort":"minimal","score":0.015,"percent":1.5,"cost":0.0015,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_external.csv:245","model":"Llama 4 Scout","version":"Llama-4-Scout-17B-16E-Instruct","effort":null,"score":0.005,"percent":0.5,"cost":0.0041,"seconds":null,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"arc_agi_external.csv:246","model":"GPT-4.1 nano","version":"gpt-4.1-nano-2025-04-14","effort":null,"score":0.0,"percent":0.0,"cost":0.0021,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"}],"cost_label":"Cost per task","time_label":null},{"id":"weirdml","name":"WeirdML","source":"https://htihle.github.io/weirdml.html","rows":[{"id":"weirdml_external.csv:0","model":"Claude Fable 5.1","version":"claude-fable-5-1_max","effort":"max","score":0.929,"percent":92.9,"cost":11.686583529411763,"seconds":null,"provider":null,"date":"2026-09-01","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"id":"weirdml_external.csv:1","model":"GPT-6 Astra","version":"gpt-6-astra_high","effort":"high","score":0.9287000000000001,"percent":92.87,"cost":2.975405612244898,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"id":"weirdml_external.csv:2","model":"Claude Fable 5.1","version":"claude-fable-5-1_high","effort":"high","score":0.923,"percent":92.30000000000001,"cost":3.2013411235955056,"seconds":null,"provider":null,"date":"2026-09-01","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"id":"weirdml_external.csv:3","model":"Claude Fable 5","version":"claude-fable-5_max","effort":"max","score":0.9194,"percent":91.94,"cost":8.424409705882354,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"weirdml_external.csv:4","model":"Claude Opus 5","version":"claude-opus-5_max","effort":"max","score":0.9178000000000001,"percent":91.78,"cost":4.7419545588235295,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"weirdml_external.csv:5","model":"Claude Opus 5","version":"claude-opus-5_high","effort":"high","score":0.9159,"percent":91.59,"cost":2.1933491176470588,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"weirdml_external.csv:6","model":"GPT-5.6 Sol (pro, max)","version":"gpt-5.6-sol_promax","effort":null,"score":0.8943000000000001,"percent":89.43,"cost":3.8707419117647057,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:7","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_high","effort":"high","score":0.8876000000000001,"percent":88.76,"cost":1.25479875,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"weirdml_external.csv:8","model":"Claude Fable 5","version":"claude-fable-5_high","effort":"high","score":0.8785,"percent":87.85,"cost":3.1068966666666666,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"weirdml_external.csv:9","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_max","effort":"max","score":0.8697,"percent":86.97,"cost":2.3122075294117646,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"weirdml_external.csv:10","model":"Claude Opus 5","version":"claude-opus-5","effort":null,"score":0.8634000000000001,"percent":86.34,"cost":1.419572117647059,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"weirdml_external.csv:11","model":"GPT-5.5","version":"gpt-5.5_xhigh","effort":"xhigh","score":0.8491,"percent":84.91,"cost":2.568345317647059,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"weirdml_external.csv:12","model":"GPT-5.5","version":"gpt-5.5_high","effort":"high","score":0.8390000000000001,"percent":83.9,"cost":2.221576529411765,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"weirdml_external.csv:13","model":"Claude Opus 4.8","version":"claude-opus-4-8_xhigh","effort":null,"score":0.8289,"percent":82.89,"cost":2.3498122772277226,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"weirdml_external.csv:14","model":"Kimi K3","version":"kimi-k3_max","effort":"max","score":0.8257,"percent":82.57,"cost":2.0536589833333334,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"weirdml_external.csv:15","model":"GPT-5.3 Codex","version":"gpt-5.3-codex","effort":null,"score":0.793,"percent":79.3,"cost":2.35,"seconds":null,"provider":null,"date":"2026-02-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"id":"weirdml_external.csv:16","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_high","effort":"high","score":0.7827,"percent":78.27,"cost":0.3119056164215686,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"weirdml_external.csv:17","model":"Claude Opus 4.6","version":"claude-opus-4-6_high","effort":"high","score":0.7795000000000001,"percent":77.95,"cost":5.779370588235294,"seconds":null,"provider":null,"date":"2026-02-05","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"id":"weirdml_external.csv:18","model":"Claude Opus 4.6","version":"claude-opus-4-6_unknown","effort":"unknown thinking","score":0.779,"percent":77.9,"cost":5.78,"seconds":null,"provider":null,"date":"2026-02-05","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"id":"weirdml_external.csv:19","model":"GPT-5.3 Codex","version":"gpt-5.3-codex_xhigh","effort":"xhigh","score":0.779,"percent":77.9,"cost":2.452577288823529,"seconds":null,"provider":null,"date":"2026-02-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"id":"weirdml_external.csv:20","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_xhigh","effort":"xhigh","score":0.777,"percent":77.7,"cost":5.719939852941176,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"weirdml_external.csv:21","model":"Claude Opus 4.7","version":"claude-opus-4-7_high","effort":"high","score":0.7644,"percent":76.44,"cost":1.8207113690476193,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"weirdml_external.csv:22","model":"Claude Opus 4.7 (no thinking)","version":"claude-opus-4-7","effort":null,"score":0.7640000000000001,"percent":76.4,"cost":0.6895180000000002,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:23","model":"Claude Opus 4.7","version":"claude-opus-4-7_unknown","effort":"unknown","score":0.764,"percent":76.4,"cost":0.6895,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"weirdml_external.csv:24","model":"Claude Opus 4.8","version":"claude-opus-4-8_medium","effort":null,"score":0.7604000000000001,"percent":76.04,"cost":0.8472938235294117,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"weirdml_external.csv:25","model":"Claude Opus 4.7","version":"claude-opus-4-7_max","effort":"max","score":0.7545000000000001,"percent":75.45,"cost":7.952621764705882,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"weirdml_external.csv:26","model":"GLM-5.3","version":"glm-5.3_max","effort":"max","score":0.7540000000000001,"percent":75.4,"cost":0.5351449411428572,"seconds":null,"provider":null,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"id":"weirdml_external.csv:27","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_xhigh","effort":"xhigh","score":0.7219,"percent":72.19,"cost":2.0473807117647063,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"weirdml_external.csv:28","model":"Gemini 3.1 Pro Preview","version":"gemini-3.1-pro-preview","effort":null,"score":0.7206999999999999,"percent":72.07,"cost":1.3601870235294122,"seconds":null,"provider":null,"date":"2026-02-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"id":"weirdml_external.csv:29","model":"Claude Opus 4.8","version":"claude-opus-4-8_none","effort":null,"score":0.7045,"percent":70.45,"cost":0.4752000990099009,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"weirdml_external.csv:30","model":"GLM-5.2","version":"glm-5.2_max","effort":null,"score":0.7012,"percent":70.12,"cost":0.5072874972,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"weirdml_external.csv:31","model":"Gemini 3 Pro Preview","version":"gemini-3-pro-preview","effort":null,"score":0.6993,"percent":69.93,"cost":0.5257711076923077,"seconds":null,"provider":null,"date":"2025-11-18","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"id":"weirdml_external.csv:32","model":"Claude Sonnet 5","version":"claude-sonnet-5_high","effort":"high","score":0.6878000000000001,"percent":68.78,"cost":0.895463294117647,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"weirdml_external.csv:33","model":"GLM-5.2","version":"glm-5.2_high","effort":null,"score":0.6731,"percent":67.31,"cost":0.28333104480874316,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"weirdml_external.csv:34","model":"Grok 4.6","version":"grok-4.6_high","effort":"high","score":0.6729,"percent":67.29,"cost":0.33483234117647054,"seconds":null,"provider":null,"date":"2026-08-12","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"id":"weirdml_external.csv:35","model":"GPT-5.5 (no thinking)","version":"gpt-5.5_none","effort":null,"score":0.6715000000000001,"percent":67.15,"cost":0.5389514705882353,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:36","model":"DeepSeek V4 Pro 0813","version":"deepseek-v4-pro-0813_max","effort":"max","score":0.6622,"percent":66.22,"cost":0.19103203769491522,"seconds":null,"provider":null,"date":"2026-08-13","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"id":"weirdml_external.csv:37","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_medium","effort":"medium","score":0.6607,"percent":66.07,"cost":2.8593499285714286,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"weirdml_external.csv:38","model":"Claude Opus 4.6 (no thinking)","version":"claude-opus-4-6","effort":null,"score":0.6587000000000001,"percent":65.87,"cost":0.7355440860215054,"seconds":null,"provider":null,"date":"2026-02-05","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:39","model":"Claude Opus 4.5 (16k thinking)","version":"claude-opus-4-5-20251101_16K","effort":null,"score":0.6374000000000001,"percent":63.74000000000001,"cost":0.7383040804597701,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:40","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_medium","effort":"medium","score":0.6344,"percent":63.44,"cost":0.4159760980392157,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"weirdml_external.csv:41","model":"DeepSeek V4 Flash 0731","version":"deepseek-v4-flash-0731_max","effort":"max","score":0.6296,"percent":62.96000000000001,"cost":0.028578264316831685,"seconds":null,"provider":null,"date":"2026-07-31","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"id":"weirdml_external.csv:42","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_high","effort":null,"score":0.6264000000000001,"percent":62.64000000000001,"cost":0.7560409661764705,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"weirdml_external.csv:43","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":null,"score":0.616,"percent":61.6,"cost":0.22158831288659797,"seconds":null,"provider":null,"date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"weirdml_external.csv:44","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_high","effort":"high","score":0.6086,"percent":60.86,"cost":0.19999024607843144,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"weirdml_external.csv:45","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_high","effort":"high","score":0.6077,"percent":60.77,"cost":0.6913346831683168,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"weirdml_external.csv:46","model":"GPT-5","version":"gpt-5-2025-08-07_high","effort":"high","score":0.6070000000000001,"percent":60.70000000000001,"cost":0.5931804044117648,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"weirdml_external.csv:47","model":"GPT-5 Pro","version":"gpt-5-pro-2025-10-06_high","effort":null,"score":0.6039,"percent":60.39,"cost":7.647644117647059,"seconds":null,"provider":null,"date":"2025-10-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"id":"weirdml_external.csv:48","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_high","effort":"high","score":0.603,"percent":60.3,"cost":1.2910174676470587,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"weirdml_external.csv:49","model":"Muse Spark 1.2","version":"muse-spark-1.2_xhigh","effort":"xhigh","score":0.603,"percent":60.3,"cost":0.38010874811320755,"seconds":null,"provider":null,"date":"2026-08-05","organization":"Meta AI","harness":"","revision":null,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"id":"weirdml_external.csv:50","model":"o3-pro","version":"o3-pro-2025-06-10_high","effort":null,"score":0.5821000000000001,"percent":58.21000000000001,"cost":2.540382702702703,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"weirdml_external.csv:51","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_none","effort":"none","score":0.5744,"percent":57.440000000000005,"cost":0.26790913235294117,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"weirdml_external.csv:52","model":"GPT-5.4 Pro (no thinking)","version":"gpt-5.4-pro-2026-03-05_none","effort":null,"score":0.5744,"percent":57.440000000000005,"cost":0.26790913235294117,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:53","model":"GLM-5.1","version":"glm-5.1","effort":null,"score":0.5710000000000001,"percent":57.10000000000001,"cost":0.3175119256097561,"seconds":null,"provider":null,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"id":"weirdml_external.csv:54","model":"DeepSeek V4 Flash 0731","version":"deepseek-v4-flash-0731_high","effort":"high","score":0.5705,"percent":57.05,"cost":0.025342430526912183,"seconds":null,"provider":null,"date":"2026-07-31","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"id":"weirdml_external.csv:55","model":"Gemini 3.6 Flash","version":"gemini-3.6-flash_high","effort":null,"score":0.561,"percent":56.10000000000001,"cost":0.6227120676470587,"seconds":null,"provider":null,"date":"2026-07-21","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"id":"weirdml_external.csv:56","model":"Kimi K2.6","version":"kimi-k2.6","effort":null,"score":0.5586,"percent":55.86,"cost":0.35035446042016805,"seconds":null,"provider":null,"date":"2026-04-20","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"id":"weirdml_external.csv:57","model":"GPT-5-codex","version":"gpt-5-codex","effort":null,"score":0.5453,"percent":54.53,"cost":1.095613721,"seconds":null,"provider":null,"date":"2025-09-15","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-codex.toml"},{"id":"weirdml_external.csv:58","model":"GPT-5-codex","version":"gpt-5-codex_high","effort":"high","score":0.5453,"percent":54.53,"cost":1.0956137205882353,"seconds":null,"provider":null,"date":"2025-09-15","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-codex.toml"},{"id":"weirdml_external.csv:59","model":"Kimi K2.7 Code","version":"kimi-k2.7-code","effort":null,"score":0.5412,"percent":54.120000000000005,"cost":0.23163436923809524,"seconds":null,"provider":null,"date":"2026-06-12","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"id":"weirdml_external.csv:60","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_16K","effort":null,"score":0.5403,"percent":54.03,"cost":0.4096601321,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:61","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_high","effort":"high","score":0.5267000000000001,"percent":52.67000000000001,"cost":0.10976651117647059,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"weirdml_external.csv:62","model":"o4-mini","version":"o4-mini-2025-04-16_high","effort":"high","score":0.5256000000000001,"percent":52.56000000000001,"cost":0.19568510000000003,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"weirdml_external.csv:63","model":"o3","version":"o3-2025-04-16_high","effort":"high","score":0.5242,"percent":52.42,"cost":0.22606690322580647,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"weirdml_external.csv:64","model":"Grok 4.20","version":"grok-4-20","effort":null,"score":0.5226,"percent":52.26,"cost":0.1262829254901961,"seconds":null,"provider":null,"date":"2026-02-17","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:65","model":"Gemma 4 31B IT","version":"gemma-4-31b-it","effort":null,"score":0.5226,"percent":52.26,"cost":0.013264671000000004,"seconds":null,"provider":null,"date":"2026-04-02","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"id":"weirdml_external.csv:66","model":"Gemini 3.1 Flash-Lite","version":"gemini-3.1-flash-lite","effort":null,"score":0.5219,"percent":52.190000000000005,"cost":0.0664239769607843,"seconds":null,"provider":null,"date":"2026-03-03","organization":"Google","harness":"","revision":null,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"id":"weirdml_external.csv:67","model":"Grok 4.3 Beta","version":"grok-4.3_unknown","effort":null,"score":0.4989,"percent":49.89,"cost":0.09031993548387096,"seconds":null,"provider":null,"date":"2026-04-17","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:68","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_none","effort":"none","score":0.49600000000000005,"percent":49.60000000000001,"cost":0.21462569999999997,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"weirdml_external.csv:69","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_low","effort":"low","score":0.496,"percent":49.6,"cost":0.215,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"weirdml_external.csv:70","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_high","effort":"high","score":0.49229999999999996,"percent":49.23,"cost":0.047747187745098035,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"weirdml_external.csv:71","model":"DeepSeek v4 Pro","version":"deepseek-v4-pro_max","effort":"max","score":0.489,"percent":48.9,"cost":0.34030045166666667,"seconds":null,"provider":null,"date":"2026-04-24","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"id":"weirdml_external.csv:72","model":"gpt-oss-120b","version":"gpt-oss-120b_high","effort":null,"score":0.4817,"percent":48.17,"cost":0.033483015,"seconds":null,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"id":"weirdml_external.csv:73","model":"gpt-oss-120b","version":"openai/gpt-oss-120b_high","effort":null,"score":0.4817,"percent":48.17,"cost":0.03348301499999999,"seconds":null,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"id":"weirdml_external.csv:74","model":"GLM-5","version":"glm-5","effort":null,"score":0.4817,"percent":48.17,"cost":0.07874515247524752,"seconds":null,"provider":null,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"id":"weirdml_external.csv:75","model":"Claude Sonnet 4.5 (16k thinking)","version":"claude-sonnet-4-5-20250929_16K","effort":null,"score":0.4771,"percent":47.71,"cost":0.3343779683544304,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:76","model":"o1-preview","version":"o1-preview-2024-09-12","effort":null,"score":0.4756,"percent":47.56,"cost":1.8487597058823528,"seconds":null,"provider":null,"date":"2024-09-12","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:77","model":"DeepSeek-V3.2-Speciale","version":"DeepSeek-V3.2-Speciale","effort":null,"score":0.4673,"percent":46.73,"cost":0.0352105525490196,"seconds":null,"provider":null,"date":"2025-12-01","organization":"DeepSeek","harness":"","revision":null},{"id":"weirdml_external.csv:78","model":"Claude Sonnet 4.5 (no thinking)","version":"claude-sonnet-4-5-20250929","effort":null,"score":0.467,"percent":46.7,"cost":0.2796832978723404,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:79","model":"DeepSeek v4 Pro","version":"deepseek-v4-pro_high","effort":"high","score":0.46530000000000005,"percent":46.53,"cost":0.12994430126582276,"seconds":null,"provider":null,"date":"2026-04-24","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"id":"weirdml_external.csv:80","model":"Grok 4.5","version":"grok-4.5_unknown","effort":"unknown thinking","score":0.4643,"percent":46.43,"cost":0.08046285647058822,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"weirdml_external.csv:81","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_16K","effort":null,"score":0.4611,"percent":46.11,"cost":0.33598588636363635,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"weirdml_external.csv:82","model":"o1","version":"o1-2024-12-17_high","effort":"high","score":0.4608,"percent":46.08,"cost":2.0286688235294115,"seconds":null,"provider":null,"date":"2024-12-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"id":"weirdml_external.csv:83","model":"Claude Opus 4.1","version":"claude-opus-4-1-20250805_16K","effort":null,"score":0.4586,"percent":45.86,"cost":1.755552705882353,"seconds":null,"provider":null,"date":"2025-08-05","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"id":"weirdml_external.csv:84","model":"Grok 4","version":"grok-4-0709","effort":null,"score":0.4573,"percent":45.73,"cost":0.4517731019417476,"seconds":null,"provider":null,"date":"2025-07-09","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:85","model":"DeepSeek v4 Flash","version":"deepseek-v4-flash_max","effort":"max","score":0.45630000000000004,"percent":45.63,"cost":0.014616010092615771,"seconds":null,"provider":null,"date":"2026-04-24","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"id":"weirdml_external.csv:86","model":"Kimi K2.5","version":"kimi-k2.5","effort":null,"score":0.456,"percent":45.6,"cost":0.052612491000000004,"seconds":null,"provider":null,"date":"2026-01-27","organization":"Moonshot","harness":"","revision":null},{"id":"weirdml_external.csv:87","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001","effort":null,"score":0.454,"percent":45.4,"cost":0.08938744000000001,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"weirdml_external.csv:88","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_16K","effort":null,"score":0.4415,"percent":44.15,"cost":0.12278560784313723,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"weirdml_external.csv:89","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514","effort":null,"score":0.4388,"percent":43.88,"cost":0.30805649999999996,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"weirdml_external.csv:90","model":"DeepSeek v4 Flash","version":"deepseek-v4-flash_high","effort":"high","score":0.4376,"percent":43.76,"cost":0.007279538164705883,"seconds":null,"provider":null,"date":"2026-04-24","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"id":"weirdml_external.csv:91","model":"Claude Opus 4","version":"claude-opus-4-20250514_16K","effort":null,"score":0.4372,"percent":43.72,"cost":1.7352728823529409,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"weirdml_external.csv:92","model":"Mistral Medium 3.5","version":"mistral-medium-2604","effort":null,"score":0.43710000000000004,"percent":43.71000000000001,"cost":0.278465325,"seconds":null,"provider":null,"date":"2026-04-28","organization":"Mistral AI","harness":"","revision":null,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"id":"weirdml_external.csv:93","model":"o3-mini","version":"o3-mini-2025-01-31_high","effort":"high","score":0.43700000000000006,"percent":43.7,"cost":0.17446746058823528,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"weirdml_external.csv:94","model":"Nemotron 3 Ultra","version":"nemotron-3-ultra","effort":null,"score":0.43450000000000005,"percent":43.45,"cost":0.0772020119760479,"seconds":null,"provider":null,"date":"2026-06-04","organization":"NVIDIA","harness":"","revision":null},{"id":"weirdml_external.csv:95","model":"Mercury 2","version":"mercury-2","effort":null,"score":0.43200000000000005,"percent":43.2,"cost":0.02977649828431373,"seconds":null,"provider":null,"date":null,"organization":"Inception Labs","harness":"","revision":null},{"id":"weirdml_external.csv:96","model":"Grok 4 Fast","version":"grok-4-fast","effort":null,"score":0.4286,"percent":42.86,"cost":0.012726230392156862,"seconds":null,"provider":null,"date":"2025-09-19","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:97","model":"Kimi K2 Thinking","version":"kimi-k2-thinking","effort":null,"score":0.4279,"percent":42.79,"cost":0.08093773510638298,"seconds":null,"provider":null,"date":"2025-11-06","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"id":"weirdml_external.csv:98","model":"Grok 3 mini","version":"grok-3-mini_high","effort":null,"score":0.4258,"percent":42.58,"cost":0.01739341941,"seconds":null,"provider":null,"date":"2025-06-24","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:99","model":"Grok-3 mini","version":"grok-3-mini-beta_high","effort":null,"score":0.4258,"percent":42.58,"cost":0.01739341941176471,"seconds":null,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:100","model":"Gemini 2.5 Flash (Sep 2025)","version":"gemini-2.5-flash-preview-09-2025_16k","effort":null,"score":0.4191,"percent":41.91,"cost":0.07324502983193276,"seconds":null,"provider":null,"date":"2025-09-25","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:101","model":"gpt-oss-120b","version":"gpt-oss-120b_medium","effort":null,"score":0.41880000000000006,"percent":41.88000000000001,"cost":0.011485011176470588,"seconds":null,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"id":"weirdml_external.csv:102","model":"DeepSeek-R1 (May 2025)","version":"DeepSeek-R1-0528","effort":null,"score":0.41630000000000006,"percent":41.63,"cost":0.07106829941176471,"seconds":null,"provider":null,"date":"2025-05-28","organization":"DeepSeek","harness":"","revision":null},{"id":"weirdml_external.csv:103","model":"Qwen3-Coder-480B-A35B","version":"Qwen3-Coder-480B-A35B-Instruct","effort":null,"score":0.4117,"percent":41.17,"cost":0.07259235294117648,"seconds":null,"provider":null,"date":"2025-07-31","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:104","model":"Qwen3-235B-A22B-Thinking (Jul 2025)","version":"Qwen3-235B-A22B-Thinking-2507","effort":null,"score":0.4104,"percent":41.04,"cost":0.01989473141,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:105","model":"Gemini 2.5 Flash (Apr 2025)","version":"gemini-2.5-flash-preview-04-17 (16K thinking)","effort":null,"score":0.40950000000000003,"percent":40.95,"cost":0.10585353676470587,"seconds":null,"provider":null,"date":"2025-04-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:106","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_16K","effort":null,"score":0.4095,"percent":40.949999999999996,"cost":0.1058535368,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:107","model":"gpt-oss-20b","version":"gpt-oss-20b_high","effort":null,"score":0.4093,"percent":40.93,"cost":0.010800342499999999,"seconds":null,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"},{"id":"weirdml_external.csv:108","model":"GLM-4.5 Thinking","version":"glm-4.5_thinking","effort":null,"score":0.4058,"percent":40.58,"cost":0.07348009059999999,"seconds":null,"provider":null,"date":"2025-08-05","organization":"Z.ai (Zhipu AI),Tsinghua University","harness":"","revision":null},{"id":"weirdml_external.csv:109","model":"Claude 3.5 Sonnet (Oct 2024)","version":"claude-3-5-sonnet-20241022","effort":null,"score":0.3997,"percent":39.97,"cost":0.24832379999999996,"seconds":null,"provider":null,"date":"2024-10-22","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:110","model":"GPT-5","version":"gpt-5-chat","effort":null,"score":0.39770000000000005,"percent":39.77,"cost":0.0785561838235294,"seconds":null,"provider":null,"date":null,"organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"weirdml_external.csv:111","model":"Qwen3.5 27B","version":"qwen3.5-27B","effort":null,"score":0.39530000000000004,"percent":39.53,"cost":0.040395000000000014,"seconds":null,"provider":null,"date":"2026-02-24","organization":"Alibaba","harness":"","revision":null,"model_id":"alibaba/qwen3.5-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-27b.toml"},{"id":"weirdml_external.csv:112","model":"DeepSeek-V3.2-Exp","version":"DeepSeek-V3.2-Exp_thinking","effort":null,"score":0.3948,"percent":39.48,"cost":0.01772299089108911,"seconds":null,"provider":null,"date":"2025-09-29","organization":"DeepSeek","harness":"","revision":null},{"id":"weirdml_external.csv:113","model":"GPT-4.5 Preview (Feb 2025)","version":"gpt-4.5-preview-2025-02-27","effort":null,"score":0.3937,"percent":39.37,"cost":1.575002205882353,"seconds":null,"provider":null,"date":"2025-02-27","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:114","model":"Kimi K2 Instruct","version":"Kimi-K2-Instruct","effort":null,"score":0.3936,"percent":39.36,"cost":0.03429974068,"seconds":null,"provider":null,"date":"2025-07-12","organization":"Moonshot","harness":"","revision":null},{"id":"weirdml_external.csv:115","model":"Kimi K2 (Jul 2025)","version":"kimi-k2-0711-preview","effort":null,"score":0.3936,"percent":39.36,"cost":0.03429974068181818,"seconds":null,"provider":null,"date":"2025-07-12","organization":"Moonshot","harness":"","revision":null},{"id":"weirdml_external.csv:116","model":"GPT-4.1","version":"gpt-4.1-2025-04-14","effort":null,"score":0.3904,"percent":39.04,"cost":0.09702961904761905,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"id":"weirdml_external.csv:117","model":"Gemini 3.5 Flash-Lite","version":"gemini-3.5-flash-lite_high","effort":null,"score":0.39,"percent":39.0,"cost":0.04675567557894736,"seconds":null,"provider":null,"date":"2026-07-21","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"id":"weirdml_external.csv:118","model":"Qwen3-235B-A22B-Instruct (Jul 2025)","version":"Qwen3-235B-A22B-Instruct-2507","effort":null,"score":0.387,"percent":38.7,"cost":0.01340708465,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:119","model":"Qwen3-235B-A22B-Thinking-2507","version":"qwen/qwen3-235b-a22b-thinking-2507","effort":null,"score":0.387,"percent":38.7,"cost":0.013407084653465346,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:120","model":"DeepSeek-V3.1","version":"DeepSeek-V3.1_thinking","effort":null,"score":0.3836,"percent":38.36,"cost":0.0311958091954023,"seconds":null,"provider":null,"date":"2025-08-21","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"id":"weirdml_external.csv:121","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_high","effort":"high","score":0.38060000000000005,"percent":38.06,"cost":0.026220323529411768,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"weirdml_external.csv:122","model":"Nemotron 3 Super","version":"nemotron-3-super","effort":null,"score":0.3801,"percent":38.01,"cost":0.012169454999999997,"seconds":null,"provider":null,"date":"2026-03-11","organization":"NVIDIA","harness":"","revision":null},{"id":"weirdml_external.csv:123","model":"GPT-5.4 nano (no thinking)","version":"gpt-5.4-nano-2026-03-17_none","effort":null,"score":0.3796,"percent":37.96,"cost":0.019951820588235296,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:124","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_none","effort":"none","score":0.379,"percent":37.9,"cost":0.07285873235294117,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"weirdml_external.csv:125","model":"GPT-4.1 mini","version":"gpt-4.1-mini-2025-04-14","effort":null,"score":0.3761,"percent":37.61,"cost":0.016606658823529415,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"id":"weirdml_external.csv:126","model":"DeepSeek-V3.1","version":"DeepSeek-V3.1","effort":null,"score":0.375,"percent":37.5,"cost":0.019730824705882354,"seconds":null,"provider":null,"date":"2025-08-21","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"id":"weirdml_external.csv:127","model":"Qwen3-235B-A22B-Thinking (Jul 2025)","version":"Qwen3-235B-A22B-Thinking-2507","effort":null,"score":0.3728,"percent":37.28,"cost":0.02087528789,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:128","model":"Qwen3-235B-A22B","version":"Qwen3-235B-A22B","effort":null,"score":0.3728,"percent":37.28,"cost":0.020875287894736843,"seconds":null,"provider":null,"date":"2025-04-29","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:129","model":"Grok 3","version":"grok-3","effort":null,"score":0.3724,"percent":37.24,"cost":0.24772464411764714,"seconds":null,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:130","model":"MiniMax-M2.7","version":"MiniMax-M2.7","effort":null,"score":0.36950000000000005,"percent":36.95,"cost":0.03490697524752475,"seconds":null,"provider":null,"date":"2026-03-18","organization":"MiniMax","harness":"","revision":null},{"id":"weirdml_external.csv:131","model":"gpt-oss-20b","version":"gpt-oss-20b_medium","effort":null,"score":0.3683,"percent":36.83,"cost":0.0038062152941176477,"seconds":null,"provider":null,"date":"2025-08-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"},{"id":"weirdml_external.csv:132","model":"Kimi K2 (Sep 2025)","version":"Kimi-K2-Instruct-0905","effort":null,"score":0.3668,"percent":36.68,"cost":0.04181978350515465,"seconds":null,"provider":null,"date":"2025-09-05","organization":"Moonshot","harness":"","revision":null},{"id":"weirdml_external.csv:133","model":"DeepSeek-R1","version":"DeepSeek-R1","effort":null,"score":0.3649,"percent":36.49,"cost":0.07850552000000001,"seconds":null,"provider":null,"date":"2025-01-20","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"id":"weirdml_external.csv:134","model":"o1-mini","version":"o1-mini-2024-09-12_medium","effort":"medium","score":0.3632,"percent":36.32,"cost":0.11127495238095236,"seconds":null,"provider":null,"date":"2024-09-12","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:135","model":"DeepSeek-V3 (Mar 2025)","version":"DeepSeek-V3-0324","effort":null,"score":0.3608,"percent":36.08,"cost":0.013903999058823524,"seconds":null,"provider":null,"date":"2025-03-24","organization":"DeepSeek","harness":"","revision":null},{"id":"weirdml_external.csv:136","model":"Gemini 2.5 Flash-Lite (Jun 2025)","version":"gemini-2.5-flash-lite-preview-06-17_16K","effort":null,"score":0.3522,"percent":35.22,"cost":0.03025316353,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:137","model":"gemini-2.5-flash-lite-preview-06-17 (Default thinking length)","version":"gemini-2.5-flash-lite-preview-06-17-thinking","effort":null,"score":0.3522,"percent":35.22,"cost":0.030253163529411764,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:138","model":"Gemma 4 26B A4B","version":"gemma-4-26b-a4b","effort":null,"score":0.3517,"percent":35.17,"cost":0.020793745,"seconds":null,"provider":null,"date":"2026-04-02","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:139","model":"grok-code-fast-1","version":"grok-code-fast-1","effort":null,"score":0.3506,"percent":35.06,"cost":0.023491821454545456,"seconds":null,"provider":null,"date":"2025-08-28","organization":null,"harness":"","revision":null},{"id":"weirdml_external.csv:140","model":"Qwen3.6 35B-A3B","version":"qwen3.6-35b-a3b","effort":null,"score":0.34490000000000004,"percent":34.49,"cost":0.05961612000000001,"seconds":null,"provider":null,"date":"2026-04-14","organization":"Alibaba","harness":"","revision":null,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"id":"weirdml_external.csv:141","model":"qwen3-coder-next","version":"qwen3-coder-next","effort":null,"score":0.344,"percent":34.4,"cost":0.032830421568627456,"seconds":null,"provider":null,"date":"2026-02-02","organization":null,"harness":"","revision":null,"model_id":"alibaba/qwen3-coder-next","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-coder-next.toml"},{"id":"weirdml_external.csv:142","model":"Mistral Medium 3.1","version":"mistral-medium-2508","effort":null,"score":0.33130000000000004,"percent":33.13,"cost":0.040179830588235294,"seconds":null,"provider":null,"date":"2025-08-12","organization":"Mistral AI","harness":"","revision":null},{"id":"weirdml_external.csv:143","model":"Inkling","version":"Inkling_high","effort":"high","score":0.3229,"percent":32.29,"cost":0.12554279170588237,"seconds":null,"provider":null,"date":"2026-07-15","organization":"Thinking Machines","harness":"","revision":null},{"id":"weirdml_external.csv:144","model":"Claude 3.5 Sonnet (Jun 2024)","version":"claude-3-5-sonnet-20240620","effort":null,"score":0.3097,"percent":30.97,"cost":0.24560110588235293,"seconds":null,"provider":null,"date":"2024-06-20","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:145","model":"Claude 3.5 Haiku (Oct 2024)","version":"claude-3-5-haiku-20241022","effort":null,"score":0.3073,"percent":30.73,"cost":0.05748017882352941,"seconds":null,"provider":null,"date":"2024-10-22","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:146","model":"Qwen3-30B-A3B","version":"Qwen3-30B-A3B","effort":null,"score":0.2975,"percent":29.75,"cost":0.008148141553,"seconds":null,"provider":null,"date":"2025-04-29","organization":"Alibaba","harness":"","revision":null,"model_id":"alibaba/qwen3-30b-a3b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-30b-a3b.toml"},{"id":"weirdml_external.csv:147","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_low","effort":"low","score":0.2589,"percent":25.89,"cost":0.004474679117647059,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"weirdml_external.csv:148","model":"Gemini 2.0 Flash (Feb 2025)","version":"gemini-2.0-flash-001","effort":null,"score":0.2577,"percent":25.77,"cost":0.006734424117647059,"seconds":null,"provider":null,"date":"2025-02-05","organization":"Google DeepMind,Google","harness":"","revision":null},{"id":"weirdml_external.csv:149","model":"GPT-4o (Nov 2024)","version":"gpt-4o-2024-11-20","effort":null,"score":0.25120000000000003,"percent":25.120000000000005,"cost":0.0835560588235294,"seconds":null,"provider":null,"date":"2024-11-20","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:150","model":"Gemini 1.5 Flash (Sep 2024)","version":"gemini-1.5-flash-002","effort":null,"score":0.2487,"percent":24.87,"cost":0.0027842108823529408,"seconds":null,"provider":null,"date":"2024-09-24","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:151","model":"Llama 4 Maverick","version":"Llama-4-Maverick-17B-128E-Instruct","effort":null,"score":0.2447,"percent":24.47,"cost":0.006132220588235293,"seconds":null,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"weirdml_external.csv:152","model":"Grok-2 (Dec 2024)","version":"grok-2-1212","effort":null,"score":0.2224,"percent":22.24,"cost":0.11412654117647059,"seconds":null,"provider":null,"date":"2024-12-12","organization":"xAI","harness":"","revision":null},{"id":"weirdml_external.csv:153","model":"Gemini 1.5 Pro (Sept 2024)","version":"gemini-1.5-pro-002","effort":null,"score":0.222,"percent":22.2,"cost":0.05045860294117647,"seconds":null,"provider":null,"date":"2024-09-24","organization":"Google DeepMind","harness":"","revision":null},{"id":"weirdml_external.csv:154","model":"Llama 3.1-405B","version":"Llama-3.1-405B-Instruct","effort":null,"score":0.2138,"percent":21.38,"cost":0.02029766588235294,"seconds":null,"provider":null,"date":"2024-07-23","organization":"Meta AI","harness":"","revision":null},{"id":"weirdml_external.csv:155","model":"Claude 3 Opus","version":"claude-3-opus-20240229","effort":null,"score":0.19219999999999998,"percent":19.22,"cost":0.856788,"seconds":null,"provider":null,"date":"2024-02-29","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:156","model":"GPT-4.1 nano","version":"gpt-4.1-nano-2025-04-14","effort":null,"score":0.1898,"percent":18.98,"cost":0.0025817174757281555,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"id":"weirdml_external.csv:157","model":"GPT-4 Turbo (Apr 2024)","version":"gpt-4-turbo-2024-04-09","effort":null,"score":0.1801,"percent":18.01,"cost":0.32102783132530116,"seconds":null,"provider":null,"date":"2024-04-09","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:158","model":"Qwen2.5-72B","version":"qwen2.5-72b-instruct","effort":null,"score":0.1597,"percent":15.97,"cost":0.011327742138364781,"seconds":null,"provider":null,"date":"2024-09-19","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:159","model":"Llama 3.3 70B","version":"Llama-3.3-70B-Instruct","effort":null,"score":0.1444,"percent":14.44,"cost":0.0014694893882352942,"seconds":null,"provider":null,"date":"2024-12-06","organization":"Meta AI","harness":"","revision":null},{"id":"weirdml_external.csv:161","model":"GPT-4 (Jun 2023)","version":"gpt-4-0613","effort":null,"score":0.1236,"percent":12.36,"cost":0.6376224705882353,"seconds":null,"provider":null,"date":"2023-06-13","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:162","model":"GPT-4o mini","version":"gpt-4o-mini-2024-07-18","effort":null,"score":0.1176,"percent":11.76,"cost":0.005260762500000001,"seconds":null,"provider":null,"date":"2024-07-18","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"id":"weirdml_external.csv:163","model":"Qwen2-72B","version":"qwen2-72b-instruct","effort":null,"score":0.113,"percent":11.3,"cost":0.010191000000000006,"seconds":null,"provider":null,"date":"2024-06-07","organization":"Alibaba","harness":"","revision":null},{"id":"weirdml_external.csv:164","model":"Claude 3 Sonnet","version":"claude-3-sonnet-20240229","effort":null,"score":0.10160000000000001,"percent":10.16,"cost":0.17829060714285716,"seconds":null,"provider":null,"date":"2024-02-29","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:165","model":"Claude 3 Haiku","version":"claude-3-haiku-20240307","effort":null,"score":0.0984,"percent":9.84,"cost":0.014075108823529416,"seconds":null,"provider":null,"date":"2024-03-07","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:166","model":"Llama 3.1-70B","version":"Llama-3.1-70B-Instruct","effort":null,"score":0.0897,"percent":8.97,"cost":0.003222669882352941,"seconds":null,"provider":null,"date":"2024-07-23","organization":"Meta AI","harness":"","revision":null},{"id":"weirdml_external.csv:167","model":"Claude 2.1","version":"claude-2.1","effort":null,"score":0.0706,"percent":7.06,"cost":0.1724719012345679,"seconds":null,"provider":null,"date":"2023-11-21","organization":"Anthropic","harness":"","revision":null},{"id":"weirdml_external.csv:168","model":"GPT-3.5 Turbo (Jan 2024)","version":"gpt-3.5-turbo-0125","effort":null,"score":0.0348,"percent":3.4799999999999995,"cost":0.04237081395348837,"seconds":null,"provider":null,"date":"2024-01-25","organization":"OpenAI","harness":"","revision":null},{"id":"weirdml_external.csv:169","model":"Mixtral 8x22B","version":"Mixtral-8x22B-Instruct-v0.1","effort":null,"score":0.0317,"percent":3.17,"cost":0.03903073333333333,"seconds":null,"provider":null,"date":"2024-04-17","organization":"Mistral AI","harness":"","revision":null},{"id":"weirdml_external.csv:170","model":"Llama 3.1-8B","version":"Llama-3.1-8B-Instruct","effort":null,"score":0.0173,"percent":1.73,"cost":0.00041039552941176473,"seconds":null,"provider":null,"date":"2024-07-23","organization":"Meta AI","harness":"","revision":null}],"cost_label":"Cost per run","time_label":null},{"id":"arc-agi-2","name":"ARC-AGI-2","source":"https://arcprize.org/leaderboard","rows":[{"id":"arc_agi_2_external.csv:2","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_max","effort":"max","score":0.925,"percent":92.5,"cost":1.44,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_2_external.csv:6","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_xhigh","effort":"xhigh","score":0.9,"percent":90.0,"cost":1.04,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_2_external.csv:17","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_high","effort":"high","score":0.8542000000000001,"percent":85.42,"cost":0.74,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_2_external.csv:19","model":"GPT-5.5","version":"gpt-5.5_xhigh","effort":"xhigh","score":0.85,"percent":85.0,"cost":1.87,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_2_external.csv:23","model":"GPT-5.5 Pro","version":"gpt-5.5-pro_high","effort":"high","score":0.8458,"percent":84.58,"cost":10.51,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"id":"arc_agi_2_external.csv:24","model":"GPT-5.5 Pro","version":"gpt-5.5-pro_xhigh","effort":"xhigh","score":0.8416,"percent":84.16,"cost":10.76,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"id":"arc_agi_2_external.csv:26","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_max","effort":"max","score":0.839,"percent":83.89999999999999,"cost":1.09,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_2_external.csv:27","model":"GPT-5.4 Pro","version":"gpt-5.4-pro-2026-03-05_xhigh","effort":"xhigh","score":0.8333,"percent":83.33,"cost":16.41,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"id":"arc_agi_2_external.csv:28","model":"GPT-5.5","version":"gpt-5.5_high","effort":"high","score":0.8333,"percent":83.33,"cost":1.45,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_2_external.csv:32","model":"Gemini 3.1 Pro Preview","version":"gemini-3.1-pro-preview","effort":null,"score":0.771,"percent":77.10000000000001,"cost":0.9622,"seconds":null,"provider":null,"date":"2026-02-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"id":"arc_agi_2_external.csv:33","model":"Gemini 3.1 Pro Preview","version":"gemini-3.1-pro-preview","effort":null,"score":0.7708,"percent":77.08,"cost":0.9622,"seconds":null,"provider":null,"date":"2026-02-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"id":"arc_agi_2_external.csv:36","model":"Claude Opus 4.7","version":"claude-opus-4-7_max","effort":"max","score":0.758,"percent":75.8,"cost":7.43,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_2_external.csv:37","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_xhigh","effort":"xhigh","score":0.7417,"percent":74.17,"cost":0.69,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_2_external.csv:39","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_xhigh","effort":"xhigh","score":0.7395,"percent":73.95,"cost":1.52,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_2_external.csv:41","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_high","effort":null,"score":0.7208,"percent":72.08,"cost":0.8503,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"arc_agi_2_external.csv:42","model":"Claude Opus 4.8","version":"claude-opus-4-8_high","effort":null,"score":0.7208,"percent":72.08,"cost":2.7425,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_2_external.csv:43","model":"Claude Opus 4.8","version":"claude-opus-4-8_medium","effort":null,"score":0.7167,"percent":71.67,"cost":2.3896,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_2_external.csv:44","model":"GPT-5.5","version":"gpt-5.5_medium","effort":"medium","score":0.7042,"percent":70.42,"cost":0.86,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_2_external.csv:48","model":"Claude Opus 4.7","version":"claude-opus-4-7_high","effort":"high","score":0.683,"percent":68.30000000000001,"cost":3.17,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_2_external.csv:49","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_high","effort":"high","score":0.675,"percent":67.5,"cost":1.02,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_2_external.csv:50","model":"Claude Opus 4.7","version":"claude-opus-4-7_medium","effort":"medium","score":0.675,"percent":67.5,"cost":2.96,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_2_external.csv:52","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_medium","effort":"medium","score":0.6708,"percent":67.08,"cost":0.47,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_2_external.csv:53","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_high","effort":"high","score":0.6708,"percent":67.08,"cost":0.55,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_2_external.csv:55","model":"Grok 4.20","version":"grok-4-20","effort":null,"score":0.6514,"percent":65.14,"cost":0.92,"seconds":null,"provider":null,"date":"2026-02-17","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:59","model":"Claude Opus 4.8","version":"claude-opus-4-8_low","effort":null,"score":0.6222,"percent":62.22,"cost":1.6818,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"arc_agi_2_external.csv:60","model":"Claude Opus 4.7","version":"claude-opus-4-7_low","effort":"low","score":0.6208,"percent":62.08,"cost":2.38,"seconds":null,"provider":null,"date":"2026-04-16","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"arc_agi_2_external.csv:64","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_high","effort":"high","score":0.6042,"percent":60.419999999999995,"cost":2.6966,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"arc_agi_2_external.csv:65","model":"Kimi K3","version":"kimi-k3_max","effort":"max","score":0.6041666666666666,"percent":60.416666666666664,"cost":1.5926940250000003,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_2_external.csv:69","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_max","effort":"max","score":0.5954,"percent":59.540000000000006,"cost":0.67,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_2_external.csv:70","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_max","effort":"max","score":0.5833,"percent":58.330000000000005,"cost":2.7179,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"arc_agi_2_external.csv:73","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_medium","effort":"medium","score":0.5542,"percent":55.42,"cost":0.68,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_2_external.csv:74","model":"Kimi K3","version":"kimi-k3_high","effort":"high","score":0.55,"percent":55.00000000000001,"cost":0.9472229249999998,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_2_external.csv:75","model":"GPT-5.2 Pro","version":"gpt-5.2-pro-2025-12-11_high","effort":null,"score":0.5416,"percent":54.16,"cost":15.721,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"id":"arc_agi_2_external.csv:77","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_xhigh","effort":"xhigh","score":0.5291,"percent":52.910000000000004,"cost":1.8982,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_2_external.csv:80","model":"Grok 4.5","version":"grok-4.5_high","effort":"high","score":0.5264,"percent":52.64,"cost":0.78,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_2_external.csv:81","model":"Grok 4.5","version":"grok-4.5_medium","effort":"medium","score":0.5264,"percent":52.64,"cost":0.75,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_2_external.csv:83","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_xhigh","effort":"xhigh","score":0.4764,"percent":47.64,"cost":0.54,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_2_external.csv:87","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_high","effort":"high","score":0.4333,"percent":43.33,"cost":1.393,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_2_external.csv:88","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_low","effort":"low","score":0.425,"percent":42.5,"cost":0.32,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"arc_agi_2_external.csv:90","model":"GPT-5.2 Pro","version":"gpt-5.2-pro-2025-12-11_medium","effort":null,"score":0.3847,"percent":38.47,"cost":8.9928,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"id":"arc_agi_2_external.csv:91","model":"Claude Opus 4.5 (64k thinking)","version":"claude-opus-4-5-20251101_64K","effort":null,"score":0.3764,"percent":37.64,"cost":2.4,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:92","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_medium","effort":"medium","score":0.375,"percent":37.5,"cost":0.27,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_2_external.csv:93","model":"Inkling","version":"Inkling","effort":null,"score":0.3652777777777778,"percent":36.52777777777778,"cost":0.6420766277310923,"seconds":null,"provider":null,"date":"2026-07-15","organization":"Thinking Machines","harness":"","revision":null},{"id":"arc_agi_2_external.csv:95","model":"GPT-5.5","version":"gpt-5.5_low","effort":"low","score":0.3333,"percent":33.33,"cost":0.35,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"arc_agi_2_external.csv:96","model":"Grok 4.5","version":"grok-4.5_low","effort":"low","score":0.3306,"percent":33.06,"cost":0.4,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"arc_agi_2_external.csv:100","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_high","effort":"high","score":0.2931,"percent":29.310000000000002,"cost":0.29,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_2_external.csv:101","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_low","effort":"low","score":0.2917,"percent":29.17,"cost":0.27,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"arc_agi_2_external.csv:103","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_medium","effort":"medium","score":0.2667,"percent":26.669999999999998,"cost":0.7594,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_2_external.csv:104","model":"Claude Opus 4.5 (16k thinking)","version":"claude-opus-4-5-20251101_16K","effort":null,"score":0.2278,"percent":22.78,"cost":0.79,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:105","model":"GLM-5.2","version":"glm-5.2_unknown","effort":null,"score":0.22775000000000004,"percent":22.775000000000002,"cost":0.2501,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"arc_agi_2_external.csv:106","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_xhigh","effort":"xhigh","score":0.189,"percent":18.9,"cost":0.75,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_2_external.csv:108","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_low","effort":"low","score":0.1875,"percent":18.75,"cost":0.17,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"arc_agi_2_external.csv:109","model":"GPT-5 Pro","version":"gpt-5-pro-2025-10-06_unknown","effort":null,"score":0.1833,"percent":18.33,"cost":7.1432,"seconds":null,"provider":null,"date":"2025-10-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"id":"arc_agi_2_external.csv:110","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_high","effort":"high","score":0.1764,"percent":17.64,"cost":1.1704,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_2_external.csv:112","model":"Claude Opus 4.5 (8k thinking)","version":"claude-opus-4-5-20251101_8K","effort":null,"score":0.1389,"percent":13.889999999999999,"cost":0.48,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:114","model":"Claude Sonnet 4.5 (32k thinking)","version":"claude-sonnet-4-5-20250929_32K","effort":null,"score":0.1361,"percent":13.61,"cost":0.7589,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:115","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_high","effort":"high","score":0.1319,"percent":13.19,"cost":0.56,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_2_external.csv:117","model":"Kimi K3","version":"kimi-k3_low","effort":"low","score":0.12361111111111112,"percent":12.36111111111111,"cost":0.2541386750000001,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"arc_agi_2_external.csv:118","model":"Kimi K2.5","version":"kimi-k2.5","effort":null,"score":0.11810000000000001,"percent":11.81,"cost":0.28,"seconds":null,"provider":null,"date":"2026-01-27","organization":"Moonshot","harness":"","revision":null},{"id":"arc_agi_2_external.csv:120","model":"GPT-5","version":"gpt-5-2025-08-07_high","effort":"high","score":0.0986,"percent":9.86,"cost":0.7302,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_2_external.csv:121","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_low","effort":"low","score":0.0972,"percent":9.719999999999999,"cost":0.2642,"seconds":null,"provider":null,"date":"2025-12-11","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"arc_agi_2_external.csv:122","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_minimal","effort":null,"score":0.0889,"percent":8.89,"cost":0.1074,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"arc_agi_2_external.csv:123","model":"Claude Opus 4","version":"claude-opus-4-20250514_16K","effort":null,"score":0.0861,"percent":8.61,"cost":1.9284,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_2_external.csv:124","model":"Claude Opus 4.5 (no thinking)","version":"claude-opus-4-5-20251101","effort":null,"score":0.0778,"percent":7.779999999999999,"cost":0.22,"seconds":null,"provider":null,"date":"2025-11-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:125","model":"GPT-5","version":"gpt-5-2025-08-07_medium","effort":"medium","score":0.0749,"percent":7.489999999999999,"cost":0.4486,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_2_external.csv:126","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_medium","effort":"medium","score":0.0736,"percent":7.359999999999999,"cost":0.11,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_2_external.csv:127","model":"Claude Sonnet 4.5 (16k thinking)","version":"claude-sonnet-4-5-20250929_16K","effort":null,"score":0.0694,"percent":6.94,"cost":0.3499,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:128","model":"Claude Sonnet 4.5 (8k thinking)","version":"claude-sonnet-4-5-20250929_8K","effort":null,"score":0.0694,"percent":6.94,"cost":0.2349,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:129","model":"o3","version":"o3-2025-04-16_high","effort":"high","score":0.0653,"percent":6.529999999999999,"cost":0.8339,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_2_external.csv:130","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_medium","effort":"medium","score":0.0653,"percent":6.529999999999999,"cost":0.4206,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_2_external.csv:131","model":"tiny-recursion-model","version":"tiny-recursion-model","effort":null,"score":0.0625,"percent":6.25,"cost":2.1,"seconds":null,"provider":null,"date":"2025-10-06","organization":null,"harness":"","revision":null},{"id":"arc_agi_2_external.csv:132","model":"o4-mini","version":"o4-mini-2025-04-16_high","effort":"high","score":0.0611,"percent":6.11,"cost":0.856,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_2_external.csv:133","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_16K","effort":null,"score":0.0593,"percent":5.93,"cost":0.4857,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_2_external.csv:134","model":"Claude Sonnet 4.5 (1k thinking)","version":"claude-sonnet-4-5-20250929_1K","effort":null,"score":0.058300000000000005,"percent":5.83,"cost":0.1424,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:135","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_xhigh","effort":"xhigh","score":0.0569,"percent":5.6899999999999995,"cost":0.16,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_2_external.csv:137","model":"Grok 4 Fast","version":"grok-4-fast","effort":null,"score":0.05280000000000001,"percent":5.28,"cost":0.061,"seconds":null,"provider":null,"date":"2025-09-19","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:139","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_low","effort":"low","score":0.05140000000000001,"percent":5.140000000000001,"cost":0.07,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"arc_agi_2_external.csv:141","model":"o3-pro","version":"o3-pro-2025-06-10_high","effort":null,"score":0.0486,"percent":4.859999999999999,"cost":7.5516,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_2_external.csv:142","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_32K","effort":null,"score":0.0486,"percent":4.859999999999999,"cost":0.757,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:143","model":"MiniMax-M2.5","version":"MiniMax-M2.5","effort":null,"score":0.0486,"percent":4.859999999999999,"cost":0.17,"seconds":null,"provider":null,"date":"2026-02-12","organization":"MiniMax","harness":"","revision":null},{"id":"arc_agi_2_external.csv:144","model":"GLM-5","version":"glm-5","effort":null,"score":0.0486,"percent":4.859999999999999,"cost":0.27,"seconds":null,"provider":null,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","harness":"","revision":null,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"id":"arc_agi_2_external.csv:145","model":"Claude Opus 4","version":"claude-opus-4-20250514_8K","effort":null,"score":0.0452,"percent":4.52,"cost":1.1569,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_2_external.csv:146","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_high","effort":"high","score":0.0444,"percent":4.44,"cost":0.1977,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_2_external.csv:147","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_medium","effort":"medium","score":0.0444,"percent":4.44,"cost":0.33,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_2_external.csv:148","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_16K","effort":null,"score":0.0403,"percent":4.03,"cost":0.7145,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:149","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_medium","effort":"medium","score":0.0403,"percent":4.03,"cost":0.0629,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_2_external.csv:150","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_32K","effort":null,"score":0.0403,"percent":4.03,"cost":0.3766,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_2_external.csv:151","model":"DeepSeek-V3.2 (Thinking; Novita)","version":"deepseek/deepseek-v3.2","effort":null,"score":0.0403,"percent":4.03,"cost":0.12,"seconds":null,"provider":null,"date":"2025-12-01","organization":"DeepSeek","harness":"","revision":null},{"id":"arc_agi_2_external.csv:153","model":"Claude Sonnet 4.5 (no thinking)","version":"claude-sonnet-4-5-20250929","effort":null,"score":0.0375,"percent":3.75,"cost":0.1295,"seconds":null,"provider":null,"date":"2025-09-29","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:154","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_high","effort":"high","score":0.0361,"percent":3.61,"cost":0.13,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_2_external.csv:156","model":"o3-mini","version":"o3-mini-2025-01-31_high","effort":"high","score":0.0299,"percent":2.9899999999999998,"cost":0.5472,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_2_external.csv:157","model":"o3","version":"o3-2025-04-16_medium","effort":"medium","score":0.0298,"percent":2.98,"cost":0.4787,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_2_external.csv:158","model":"Gemini 2.5 Pro (Jun 2025)","version":"gemini-2.5-pro_8K","effort":null,"score":0.0292,"percent":2.92,"cost":0.4439,"seconds":null,"provider":null,"date":"2025-06-17","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:159","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_16K","effort":null,"score":0.0278,"percent":2.78,"cost":0.1392,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_2_external.csv:161","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_high","effort":"high","score":0.026100000000000005,"percent":2.6100000000000003,"cost":0.0295,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_2_external.csv:162","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_23K","effort":null,"score":0.025400000000000002,"percent":2.54,"cost":0.3191,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:163","model":"o4-mini","version":"o4-mini-2025-04-16_medium","effort":"medium","score":0.0236,"percent":2.36,"cost":0.2311,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_2_external.csv:164","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_1K","effort":null,"score":0.0216,"percent":2.16,"cost":0.0302,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:165","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_8K","effort":null,"score":0.0212,"percent":2.12,"cost":0.1994,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:166","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_8K","effort":null,"score":0.0212,"percent":2.12,"cost":0.2654,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_2_external.csv:168","model":"o3-mini","version":"o3-mini-2025-01-31_medium","effort":"medium","score":0.020800000000000003,"percent":2.08,"cost":0.2843,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_2_external.csv:169","model":"o3-pro","version":"o3-pro-2025-06-10_low","effort":null,"score":0.020500000000000004,"percent":2.0500000000000003,"cost":2.2293,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_2_external.csv:170","model":"o3","version":"o3-2025-04-16_low","effort":"low","score":0.0199,"percent":1.9900000000000002,"cost":0.2343,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"id":"arc_agi_2_external.csv:171","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20_16K","effort":null,"score":0.0198,"percent":1.9800000000000002,"cost":0.3173,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:172","model":"o3-pro","version":"o3-pro-2025-06-10_medium","effort":null,"score":0.0194,"percent":1.94,"cost":4.7441,"seconds":null,"provider":null,"date":"2025-06-10","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"id":"arc_agi_2_external.csv:173","model":"GPT-5","version":"gpt-5-2025-08-07_low","effort":"low","score":0.0194,"percent":1.94,"cost":0.1896,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_2_external.csv:174","model":"GPT-5","version":"gpt-5-2025-08-07_low","effort":"low","score":0.0194,"percent":1.94,"cost":0.1896,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_2_external.csv:175","model":"GPT-5.1","version":"gpt-5.1-2025-11-13_low","effort":"low","score":0.0194,"percent":1.94,"cost":0.129,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"id":"arc_agi_2_external.csv:176","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_medium","effort":"medium","score":0.0194,"percent":1.94,"cost":0.06,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_2_external.csv:177","model":"Gemini 2.5 Flash (May 2025)","version":"gemini-2.5-flash-preview-05-20","effort":null,"score":0.0169,"percent":1.69,"cost":0.057,"seconds":null,"provider":null,"date":"2025-05-20","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:178","model":"o4-mini","version":"o4-mini-2025-04-16_low","effort":"low","score":0.0167,"percent":1.67,"cost":0.05,"seconds":null,"provider":null,"date":"2025-04-16","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"id":"arc_agi_2_external.csv:179","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_minimal","effort":"minimal","score":0.0167,"percent":1.67,"cost":0.0094,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_2_external.csv:180","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_8K","effort":null,"score":0.0167,"percent":1.67,"cost":0.0909,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_2_external.csv:181","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_low","effort":"low","score":0.015300000000000001,"percent":1.53,"cost":0.01,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"arc_agi_2_external.csv:183","model":"Gemini 2.0 Flash (Feb 2025)","version":"gemini-2.0-flash-001","effort":null,"score":0.013000000000000001,"percent":1.3,"cost":0.004,"seconds":null,"provider":null,"date":"2025-02-05","organization":"Google DeepMind,Google","harness":"","revision":null},{"id":"arc_agi_2_external.csv:184","model":"DeepSeek-R1","version":"DeepSeek-R1","effort":null,"score":0.013000000000000001,"percent":1.3,"cost":0.08,"seconds":null,"provider":null,"date":"2025-01-20","organization":"DeepSeek","harness":"","revision":null,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"id":"arc_agi_2_external.csv:185","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514","effort":null,"score":0.012700000000000001,"percent":1.27,"cost":0.1272,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_2_external.csv:186","model":"codex-mini-2025-05-16","version":"codex-mini-2025-05-16","effort":null,"score":0.012700000000000001,"percent":1.27,"cost":0.23,"seconds":null,"provider":null,"date":"2025-05-16","organization":null,"harness":"","revision":null},{"id":"arc_agi_2_external.csv:187","model":"Claude Opus 4","version":"claude-opus-4-20250514","effort":null,"score":0.012700000000000001,"percent":1.27,"cost":0.6388,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_2_external.csv:188","model":"Qwen3-235B-A22B-Instruct (Jul 2025)","version":"Qwen3-235B-A22B-Instruct-2507","effort":null,"score":0.0125,"percent":1.25,"cost":0.0044,"seconds":null,"provider":null,"date":"2025-07-25","organization":"Alibaba","harness":"","revision":null},{"id":"arc_agi_2_external.csv:189","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001","effort":null,"score":0.0125,"percent":1.25,"cost":0.0426,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_2_external.csv:191","model":"Claude Haiku 4.5","version":"claude-haiku-4-5-20251001_1K","effort":null,"score":0.0125,"percent":1.25,"cost":0.0471,"seconds":null,"provider":null,"date":"2025-10-15","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"id":"arc_agi_2_external.csv:192","model":"DeepSeek-R1 (May 2025)","version":"DeepSeek-R1-0528","effort":null,"score":0.0112,"percent":1.1199999999999999,"cost":0.0527,"seconds":null,"provider":null,"date":"2025-05-28","organization":"DeepSeek","harness":"","revision":null},{"id":"arc_agi_2_external.csv:193","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_low","effort":"low","score":0.0111,"percent":1.11,"cost":0.06,"seconds":null,"provider":null,"date":"2026-03-17","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"arc_agi_2_external.csv:194","model":"Claude 3.7 Sonnet (8k thinking)","version":"claude-3-7-sonnet-20250219_8K","effort":null,"score":0.009,"percent":0.8999999999999999,"cost":0.36,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:195","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_medium","effort":"medium","score":0.0088,"percent":0.88,"cost":0.0137,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_2_external.csv:196","model":"Claude Sonnet 4","version":"claude-sonnet-4-20250514_1K","effort":null,"score":0.0085,"percent":0.8500000000000001,"cost":0.1425,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"id":"arc_agi_2_external.csv:199","model":"o1-mini","version":"o1-mini-2024-09-12_unknown","effort":"unknown thinking","score":0.0083,"percent":0.83,"cost":0.1907,"seconds":null,"provider":null,"date":"2024-09-12","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:201","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_low","effort":"low","score":0.0083,"percent":0.83,"cost":0.0189,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"arc_agi_2_external.csv:202","model":"GPT-4.5 Preview (Feb 2025)","version":"gpt-4.5-preview-2025-02-27","effort":null,"score":0.008,"percent":0.8,"cost":2.1,"seconds":null,"provider":null,"date":"2025-02-27","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:203","model":"Gemini 1.5 Pro (Sept 2024)","version":"gemini-1.5-pro-002","effort":null,"score":0.008,"percent":0.8,"cost":0.04,"seconds":null,"provider":null,"date":"2024-09-24","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:204","model":"Claude 3.7 Sonnet (16k thinking)","version":"claude-3-7-sonnet-20250219_16K","effort":null,"score":0.007000000000000001,"percent":0.7000000000000001,"cost":0.51,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:205","model":"GPT-4.1","version":"gpt-4.1-2025-04-14","effort":null,"score":0.0042,"percent":0.42,"cost":0.0691,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"id":"arc_agi_2_external.csv:206","model":"Grok 3 mini","version":"grok-3-mini_low","effort":null,"score":0.0042,"percent":0.42,"cost":0.0131,"seconds":null,"provider":null,"date":"2025-06-24","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:207","model":"GPT-5.1 (no thinking)","version":"gpt-5.1-2025-11-13_none","effort":null,"score":0.0042,"percent":0.42,"cost":0.0581,"seconds":null,"provider":null,"date":"2025-11-13","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:208","model":"Claude 3.7 Sonnet (1k thinking)","version":"claude-3-7-sonnet-20250219_1K","effort":null,"score":0.004,"percent":0.4,"cost":0.14,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:209","model":"Claude 3.7 Sonnet","version":"claude-3-7-sonnet-20250219","effort":null,"score":0.0,"percent":0.0,"cost":0.12,"seconds":null,"provider":null,"date":"2025-02-24","organization":"Anthropic","harness":"","revision":null},{"id":"arc_agi_2_external.csv:210","model":"GPT-4o mini","version":"gpt-4o-mini-2024-07-18","effort":null,"score":0.0,"percent":0.0,"cost":0.01,"seconds":null,"provider":null,"date":"2024-07-18","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"id":"arc_agi_2_external.csv:211","model":"GPT-4.1 nano","version":"gpt-4.1-nano-2025-04-14","effort":null,"score":0.0,"percent":0.0,"cost":0.0036,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"id":"arc_agi_2_external.csv:212","model":"Llama 4 Scout","version":"Llama-4-Scout-17B-16E-Instruct","effort":null,"score":0.0,"percent":0.0,"cost":0.0062,"seconds":null,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:213","model":"Llama 4 Maverick","version":"Llama-4-Maverick-17B-128E-Instruct","effort":null,"score":0.0,"percent":0.0,"cost":0.0121,"seconds":null,"provider":null,"date":"2025-04-05","organization":"Meta AI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:214","model":"o3-mini","version":"o3-mini-2025-01-31_low","effort":null,"score":0.0,"percent":0.0,"cost":0.0623,"seconds":null,"provider":null,"date":"2025-01-31","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"id":"arc_agi_2_external.csv:215","model":"GPT-4.1 mini","version":"gpt-4.1-mini-2025-04-14","effort":null,"score":0.0,"percent":0.0,"cost":0.0139,"seconds":null,"provider":null,"date":"2025-04-14","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"id":"arc_agi_2_external.csv:216","model":"Grok 3","version":"grok-3","effort":null,"score":0.0,"percent":0.0,"cost":0.1421,"seconds":null,"provider":null,"date":"2025-04-09","organization":"xAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:217","model":"Claude Opus 4","version":"claude-opus-4-20250514_1K","effort":null,"score":0.0,"percent":0.0,"cost":0.7503,"seconds":null,"provider":null,"date":"2025-05-22","organization":"Anthropic","harness":"","revision":null,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"id":"arc_agi_2_external.csv:218","model":"Magistral Small 1.0","version":"magistral-small-2506","effort":null,"score":0.0,"percent":0.0,"cost":0.0488,"seconds":null,"provider":null,"date":"2025-06-10","organization":"Mistral AI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:220","model":"Gemini 2.5 Pro Preview (Jun 2025)","version":"gemini-2.5-pro-preview-06-05_1K","effort":null,"score":0.0,"percent":0.0,"cost":0.0885,"seconds":null,"provider":null,"date":"2025-06-05","organization":"Google DeepMind","harness":"","revision":null},{"id":"arc_agi_2_external.csv:221","model":"GPT-5","version":"gpt-5-2025-08-07_minimal","effort":"minimal","score":0.0,"percent":0.0,"cost":0.0562,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"arc_agi_2_external.csv:222","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_low","effort":"low","score":0.0,"percent":0.0,"cost":0.0033,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"arc_agi_2_external.csv:223","model":"GPT-4o (Nov 2024)","version":"gpt-4o-2024-11-20","effort":null,"score":0.0,"percent":0.0,"cost":0.08,"seconds":null,"provider":null,"date":"2024-11-20","organization":"OpenAI","harness":"","revision":null},{"id":"arc_agi_2_external.csv:224","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_minimal","effort":"minimal","score":0.0,"percent":0.0,"cost":0.0025,"seconds":null,"provider":null,"date":"2025-08-07","organization":"OpenAI","harness":"","revision":null,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"}],"cost_label":"Cost per task","time_label":null},{"id":"deepswe-mini-swe-agent","name":"DeepSWE · mini-swe-agent","source":"https://deepswe.datacurve.ai/","rows":[{"id":"deepswe_external.csv:0","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_max","effort":"max","score":0.7266666666666667,"percent":72.66666666666667,"cost":8.386436346666667,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"deepswe_external.csv:1","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_xhigh","effort":"xhigh","score":0.7073170731707318,"percent":70.73170731707317,"cost":4.703663860310422,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"deepswe_external.csv:2","model":"Claude Fable 5","version":"claude-fable-5_xhigh","effort":"xhigh","score":0.6991150442477876,"percent":69.91150442477876,"cost":13.414521495535714,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"deepswe_external.csv:3","model":"Claude Fable 5","version":"claude-fable-5_max","effort":"max","score":0.6972477064220184,"percent":69.72477064220183,"cost":21.634702092592594,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"deepswe_external.csv:4","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_max","effort":"max","score":0.6962305986696231,"percent":69.62305986696231,"cost":4.945847263858093,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"deepswe_external.csv:5","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_high","effort":"high","score":0.6940133037694013,"percent":69.40133037694012,"cost":3.4698331108647453,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"deepswe_external.csv:6","model":"Claude Fable 5","version":"claude-fable-5_high","effort":"high","score":0.6860465116279071,"percent":68.6046511627907,"cost":9.177638214788733,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"deepswe_external.csv:7","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_max","effort":"max","score":0.671875,"percent":67.1875,"cost":3.028116810267857,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"deepswe_external.csv:8","model":"GPT-5.5","version":"gpt-5.5_xhigh","effort":"xhigh","score":0.6703539823008849,"percent":67.03539823008849,"cost":7.226236674778761,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"deepswe_external.csv:9","model":"GPT-5.5","version":"gpt-5.5_high","effort":"high","score":0.6438053097345133,"percent":64.38053097345133,"cost":5.100437004424779,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"deepswe_external.csv:10","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_medium","effort":"medium","score":0.6106194690265486,"percent":61.06194690265486,"cost":1.8620318075221238,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"deepswe_external.csv:11","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_xhigh","effort":"xhigh","score":0.6017699115044248,"percent":60.17699115044248,"cost":2.127191725663717,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"deepswe_external.csv:12","model":"Claude Opus 4.8","version":"claude-opus-4-8_max","effort":"max","score":0.5897435897435898,"percent":58.97435897435898,"cost":13.222583593240094,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"deepswe_external.csv:13","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_xhigh","effort":"xhigh","score":0.5685840707964602,"percent":56.85840707964602,"cost":1.5356192685840708,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"deepswe_external.csv:14","model":"Claude Opus 4.8","version":"claude-opus-4-8_xhigh","effort":"xhigh","score":0.5436241610738255,"percent":54.36241610738255,"cost":8.006355704138702,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"deepswe_external.csv:15","model":"GPT-5.5","version":"gpt-5.5_medium","effort":"medium","score":0.5398230088495575,"percent":53.98230088495575,"cost":2.749342734513274,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"deepswe_external.csv:16","model":"Claude Sonnet 5","version":"claude-sonnet-5_max","effort":"max","score":0.5384615384615384,"percent":53.84615384615385,"cost":26.399858950791852,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"deepswe_external.csv:17","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_high","effort":"high","score":0.5376106194690266,"percent":53.76106194690266,"cost":1.1343762765486725,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"deepswe_external.csv:18","model":"Muse Spark 1.1","version":"muse-spark-1.1","effort":"xhigh","score":0.5331858407079646,"percent":53.31858407079646,"cost":2.361142644469026,"seconds":null,"provider":null,"date":"2026-07-09","organization":"Meta AI","harness":"mini-swe-agent","revision":null,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"id":"deepswe_external.csv:19","model":"Claude Opus 4.8","version":"claude-opus-4-8_high","effort":"high","score":0.5176991150442478,"percent":51.76991150442478,"cost":4.281940017699115,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"deepswe_external.csv:20","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_xhigh","effort":"xhigh","score":0.5176991150442478,"percent":51.76991150442478,"cost":5.6524577245575225,"seconds":null,"provider":null,"date":"2026-03-05","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"deepswe_external.csv:21","model":"Claude Sonnet 5","version":"claude-sonnet-5_xhigh","effort":"xhigh","score":0.49667405764966743,"percent":49.667405764966745,"cost":11.890634540022173,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"deepswe_external.csv:22","model":"Claude Opus 4.8","version":"claude-opus-4-8_medium","effort":"medium","score":0.48672566371681414,"percent":48.67256637168141,"cost":3.4438571028761062,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"deepswe_external.csv:23","model":"Claude Sonnet 5","version":"claude-sonnet-5_high","effort":"high","score":0.4823008849557522,"percent":48.23008849557522,"cost":7.425560161197339,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"deepswe_external.csv:24","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_low","effort":"low","score":0.45353982300884954,"percent":45.35398230088495,"cost":1.0742664646017699,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"deepswe_external.csv:25","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_high","effort":"high","score":0.4424778761061947,"percent":44.24778761061947,"cost":0.7779007499999999,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"deepswe_external.csv:26","model":"GLM-5.2","version":"glm-5.2_max","effort":"max","score":0.4377777777777778,"percent":43.777777777777786,"cost":3.9199075743111114,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"mini-swe-agent","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"deepswe_external.csv:27","model":"Claude Opus 4.8","version":"claude-opus-4-8_low","effort":"low","score":0.4079822616407982,"percent":40.79822616407982,"cost":2.293369216740577,"seconds":null,"provider":null,"date":"2026-05-28","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"deepswe_external.csv:28","model":"GLM-5.2","version":"glm-5.2_high","effort":"high","score":0.36283185840707965,"percent":36.283185840707965,"cost":2.8355164725663715,"seconds":null,"provider":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"mini-swe-agent","revision":null,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"deepswe_external.csv:29","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_medium","effort":"medium","score":0.3511111111111111,"percent":35.11111111111111,"cost":0.5832493688888889,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"deepswe_external.csv:30","model":"Kimi K2.7 Code","version":"kimi-k2.7-code","effort":null,"score":0.3053097345132743,"percent":30.53097345132743,"cost":2.815536243274336,"seconds":null,"provider":null,"date":"2026-06-12","organization":"Moonshot","harness":"mini-swe-agent","revision":null,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"id":"deepswe_external.csv:31","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_high","effort":"high","score":0.29933481152993346,"percent":29.933481152993345,"cost":5.5223517987804875,"seconds":null,"provider":null,"date":"2026-02-17","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"deepswe_external.csv:32","model":"GPT-5.5","version":"gpt-5.5_low","effort":"low","score":0.26991150442477874,"percent":26.991150442477874,"cost":1.2002025442477877,"seconds":null,"provider":null,"date":"2026-04-23","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"deepswe_external.csv:33","model":"GPT-5.6 Terra","version":"gpt-5.6-terra_low","effort":"low","score":0.24053452115812918,"percent":24.053452115812917,"cost":0.4277452605790646,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"id":"deepswe_external.csv:34","model":"Gemini 3.1 Pro Preview","version":"gemini-3.1-pro-preview","effort":"high","score":0.1172566371681416,"percent":11.725663716814159,"cost":2.1434045568888886,"seconds":null,"provider":null,"date":"2026-02-19","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"id":"deepswe_external.csv:35","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_medium","effort":"medium","score":0.11283185840707964,"percent":11.283185840707963,"cost":0.21630978982300886,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"deepswe_external.csv:36","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_low","effort":"low","score":0.015486725663716814,"percent":1.5486725663716814,"cost":0.07240623938053097,"seconds":null,"provider":null,"date":"2026-07-09","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"deepswe_external.csv:37","model":"Claude Opus 5","version":"claude-opus-5_max","effort":"max","score":0.7364864864864865,"percent":73.64864864864865,"cost":11.837583271396396,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"deepswe_external.csv:38","model":"Claude Opus 5","version":"claude-opus-5_xhigh","effort":"xhigh","score":0.7315436241610739,"percent":73.15436241610739,"cost":9.07220232326622,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"deepswe_external.csv:39","model":"Claude Opus 5","version":"claude-opus-5_high","effort":"high","score":0.7282850779510023,"percent":72.82850779510022,"cost":6.076084489977728,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"deepswe_external.csv:40","model":"Claude Opus 5","version":"claude-opus-5_medium","effort":"medium","score":0.6890380313199106,"percent":68.90380313199105,"cost":3.28978764541387,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"deepswe_external.csv:41","model":"Kimi K3","version":"kimi-k3_max","effort":"max","score":0.6851441241685144,"percent":68.51441241685144,"cost":4.654682129933482,"seconds":null,"provider":null,"date":"2026-07-16","organization":"Moonshot","harness":"mini-swe-agent","revision":null,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"deepswe_external.csv:42","model":"Claude Fable 5","version":"claude-fable-5_medium","effort":"medium","score":0.6536697247706422,"percent":65.36697247706422,"cost":6.088186528935186,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"deepswe_external.csv:43","model":"Claude Fable 5","version":"claude-fable-5_low","effort":"low","score":0.5958429561200924,"percent":59.58429561200924,"cost":3.757873125874126,"seconds":null,"provider":null,"date":"2026-06-09","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"deepswe_external.csv:44","model":"Claude Opus 5","version":"claude-opus-5_low","effort":"low","score":0.5812917594654788,"percent":58.12917594654788,"cost":1.6626223919821828,"seconds":null,"provider":null,"date":"2026-07-24","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"deepswe_external.csv:45","model":"Grok 4.5","version":"grok-4.5_high","effort":"high","score":0.5376106194690266,"percent":53.76106194690266,"cost":2.4157481725663716,"seconds":null,"provider":null,"date":"2026-07-08","organization":"xAI","harness":"mini-swe-agent","revision":null,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"deepswe_external.csv:46","model":"Gemini 3.6 Flash","version":"gemini-3.6-flash_high","effort":"high","score":0.4668141592920354,"percent":46.68141592920354,"cost":4.419013277333334,"seconds":null,"provider":null,"date":"2026-07-21","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"id":"deepswe_external.csv:47","model":"Claude Sonnet 5","version":"claude-sonnet-5_medium","effort":"medium","score":0.3977777777777778,"percent":39.77777777777778,"cost":4.079036229,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"deepswe_external.csv:48","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_medium","effort":"medium","score":0.37389380530973454,"percent":37.389380530973455,"cost":7.341823985508849,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"deepswe_external.csv:49","model":"Claude Sonnet 5","version":"claude-sonnet-5_low","effort":"low","score":0.3051224944320713,"percent":30.51224944320713,"cost":2.1865550832962137,"seconds":null,"provider":null,"date":"2026-06-30","organization":"Anthropic","harness":"mini-swe-agent","revision":null,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"deepswe_external.csv:50","model":"Grok 4.6","version":"grok-4.6_medium","effort":"medium","score":0.6747787610619469,"percent":67.47787610619469,"cost":3.4489837522123894,"seconds":null,"provider":null,"date":"2026-08-12","organization":"xAI","harness":"mini-swe-agent","revision":null,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"id":"deepswe_external.csv:51","model":"Grok 4.6","version":"grok-4.6_xhigh","effort":"xhigh","score":0.6674057649667405,"percent":66.74057649667405,"cost":5.497668745011087,"seconds":null,"provider":null,"date":"2026-08-12","organization":"xAI","harness":"mini-swe-agent","revision":null,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"id":"deepswe_external.csv:52","model":"Grok 4.6","version":"grok-4.6_high","effort":"high","score":0.6518847006651884,"percent":65.18847006651885,"cost":4.3848732239467845,"seconds":null,"provider":null,"date":"2026-08-12","organization":"xAI","harness":"mini-swe-agent","revision":null,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"id":"deepswe_external.csv:53","model":"Qwen3.8 Max","version":"qwen3.8-max_xhigh","effort":"xhigh","score":0.5746102449888641,"percent":57.46102449888642,"cost":3.7290646500445437,"seconds":null,"provider":null,"date":"2026-08-02","organization":"Alibaba","harness":"mini-swe-agent","revision":null,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"id":"deepswe_external.csv:54","model":"Grok 4.6","version":"grok-4.6_low","effort":"low","score":0.41648106904231624,"percent":41.648106904231625,"cost":1.0423604231625836,"seconds":null,"provider":null,"date":"2026-08-12","organization":"xAI","harness":"mini-swe-agent","revision":null,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"id":"deepswe_external.csv:55","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_high","effort":"high","score":0.3606194690265487,"percent":36.06194690265487,"cost":3.4467114588691796,"seconds":null,"provider":null,"date":"2026-05-19","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"deepswe_external.csv:56","model":"Gemini 3.7 Flash","version":"gemini-3.7-flash_medium","effort":"medium","score":0.6548672566371682,"percent":65.48672566371681,"cost":2.0251351826880533,"seconds":null,"provider":null,"date":"2026-08-13","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"id":"deepswe_external.csv:57","model":"Gemini 3.7 Flash","version":"gemini-3.7-flash_high","effort":"high","score":0.6526548672566371,"percent":65.2654867256637,"cost":2.1762860351216813,"seconds":null,"provider":null,"date":"2026-08-13","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"id":"deepswe_external.csv:58","model":"Gemini 3.7 Flash","version":"gemini-3.7-flash_low","effort":"low","score":0.5376106194690266,"percent":53.76106194690266,"cost":1.8322645629424779,"seconds":null,"provider":null,"date":"2026-08-13","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"id":"deepswe_external.csv:59","model":"GLM-5.3","version":"glm-5.3_max","effort":"max","score":0.6895787139689579,"percent":68.95787139689578,"cost":3.9933584893126386,"seconds":null,"provider":null,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","harness":"mini-swe-agent","revision":null,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"id":"deepswe_external.csv:60","model":"Muse Spark 1.2","version":"muse-spark-1.2_xhigh","effort":"xhigh","score":0.5486725663716814,"percent":54.86725663716814,"cost":3.6955048180309733,"seconds":null,"provider":null,"date":"2026-08-05","organization":"Meta AI","harness":"mini-swe-agent","revision":null,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"id":"deepswe_external.csv:61","model":"Gemini 3.8 Flash","version":"gemini-3.8-flash_high","effort":"high","score":0.738255033557047,"percent":73.8255033557047,"cost":2.362349413758389,"seconds":null,"provider":null,"date":"2026-09-02","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"id":"deepswe_external.csv:62","model":"Gemini 3.8 Flash","version":"gemini-3.8-flash_medium","effort":"medium","score":0.7101769911504425,"percent":71.01769911504425,"cost":1.9671024789823006,"seconds":null,"provider":null,"date":"2026-09-02","organization":"Google DeepMind","harness":"mini-swe-agent","revision":null,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"id":"deepswe_external.csv:63","model":"GLM-5.3-Flash","version":"glm-5.3-flash_max","effort":"max","score":0.6339285714285714,"percent":63.39285714285714,"cost":0.48196371245535713,"seconds":null,"provider":null,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","harness":"mini-swe-agent","revision":null,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"id":"deepswe_external.csv:64","model":"GPT-6 Astra","version":"gpt-6-astra_xhigh","effort":"xhigh","score":0.7411504424778761,"percent":74.11504424778761,"cost":6.52377356460177,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"id":"deepswe_external.csv:65","model":"GPT-6 Astra","version":"gpt-6-astra_high","effort":"high","score":0.7323008849557522,"percent":73.23008849557522,"cost":5.723721956194691,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"id":"deepswe_external.csv:66","model":"GPT-6 Astra","version":"gpt-6-astra_max","effort":"max","score":0.7323008849557522,"percent":73.23008849557522,"cost":12.369032401327434,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"id":"deepswe_external.csv:67","model":"GPT-6 Astra","version":"gpt-6-astra_medium","effort":"medium","score":0.7278761061946902,"percent":72.78761061946902,"cost":4.379527705752213,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"id":"deepswe_external.csv:68","model":"GPT-6 Astra","version":"gpt-6-astra_low","effort":"low","score":0.6703539823008849,"percent":67.03539823008849,"cost":2.188762706637168,"seconds":null,"provider":null,"date":"2026-09-03","organization":"OpenAI","harness":"mini-swe-agent","revision":null,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"}],"cost_label":"Mean cost (USD)","time_label":null},{"id":"proofbench-v1-1","name":"ProofBench · v1.1","source":"https://www.vals.ai/benchmarks/proof_bench","rows":[{"id":"proofbench_external.csv:1","model":"Claude Opus 5","version":"claude-opus-5_max","effort":"max","score":0.99,"percent":99.0,"cost":1.786125,"seconds":590.9,"provider":"Anthropic","date":"2026-07-24","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"id":"proofbench_external.csv:3","model":"Claude Fable 5","version":"claude-fable-5_max","effort":"max","score":0.9500000000000001,"percent":95.0,"cost":5.24099,"seconds":824.8,"provider":"Anthropic","date":"2026-06-09","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"id":"proofbench_external.csv:4","model":"Kimi K3","version":"kimi-k3_unknown","effort":"unknown","score":0.87,"percent":87.0,"cost":1.646802,"seconds":2033.4,"provider":"Moonshot AI","date":"2026-07-16","organization":"Moonshot","harness":"","revision":"v1.1","model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"id":"proofbench_external.csv:5","model":"GPT-5.6 Sol","version":"gpt-5.6-sol_max","effort":"max","score":0.8300000000000001,"percent":83.0,"cost":1.42393,"seconds":520.4,"provider":"OpenAI","date":"2026-07-09","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"id":"proofbench_external.csv:6","model":"Claude Sonnet 5","version":"claude-sonnet-5_max","effort":"max","score":0.77,"percent":77.0,"cost":2.275898,"seconds":937.1,"provider":"Anthropic","date":"2026-06-30","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"id":"proofbench_external.csv:8","model":"Claude Opus 4.8","version":"claude-opus-4-8_max","effort":"max","score":0.6900000000000001,"percent":69.0,"cost":2.293129,"seconds":1166.322,"provider":"Anthropic","date":"2026-05-28","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"id":"proofbench_external.csv:9","model":"GPT-5.6 Luna","version":"gpt-5.6-luna_max","effort":"max","score":0.6,"percent":60.0,"cost":0.096675,"seconds":513.7,"provider":"OpenAI","date":"2026-07-09","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"id":"proofbench_external.csv:12","model":"GPT-5.4","version":"gpt-5.4-2026-03-05_xhigh","effort":"xhigh","score":0.56,"percent":56.00000000000001,"cost":3.202015,"seconds":2751.526,"provider":"OpenAI","date":"2026-03-05","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"id":"proofbench_external.csv:14","model":"Claude Opus 4.7","version":"claude-opus-4-7_max","effort":"max","score":0.54,"percent":54.0,"cost":1.807786,"seconds":679.708,"provider":"Anthropic","date":"2026-04-16","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"id":"proofbench_external.csv:16","model":"GPT-5.5","version":"gpt-5.5_xhigh","effort":"xhigh","score":0.5,"percent":50.0,"cost":1.68179,"seconds":2953.443,"provider":"OpenAI","date":"2026-04-23","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"id":"proofbench_external.csv:17","model":"Claude Opus 4.6","version":"claude-opus-4-6_max","effort":"max","score":0.5,"percent":50.0,"cost":2.703583,"seconds":3973.892,"provider":"Anthropic","date":"2026-02-05","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"id":"proofbench_external.csv:21","model":"Claude Sonnet 4.6","version":"claude-sonnet-4-6_max","effort":"max","score":0.45,"percent":45.0,"cost":2.162108,"seconds":5463.593,"provider":"Anthropic","date":"2026-02-17","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"id":"proofbench_external.csv:23","model":"Muse Spark 1.1","version":"muse-spark-1.1","effort":"xhigh","score":0.39,"percent":39.0,"cost":0.519718,"seconds":1056.902,"provider":"Meta","date":"2026-07-09","organization":"Meta AI","harness":"","revision":"v1.1","model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"id":"proofbench_external.csv:24","model":"Claude Opus 4.5","version":"claude-opus-4-5-20251101_unknown","effort":"high","score":0.36,"percent":36.0,"cost":1.267043,"seconds":3757.671,"provider":"Anthropic","date":"2025-11-24","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"id":"proofbench_external.csv:25","model":"Gemini 3.6 Flash","version":"gemini-3.6-flash_unknown","effort":"high","score":0.36,"percent":36.0,"cost":0.522059,"seconds":346.937,"provider":"Google","date":"2026-07-21","organization":"Google DeepMind","harness":"","revision":"v1.1","model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"id":"proofbench_external.csv:26","model":"GLM-5.2","version":"glm-5.2_max","effort":"max","score":0.35000000000000003,"percent":35.0,"cost":0.0,"seconds":null,"provider":"Zhipu AI","date":"2026-06-16","organization":"Z.ai (Zhipu AI)","harness":"","revision":"v1.1","model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"id":"proofbench_external.csv:27","model":"Grok 4.5","version":"grok-4.5_high","effort":"high","score":0.31,"percent":31.0,"cost":0.573991,"seconds":668.86,"provider":"SpaceXAI","date":"2026-07-08","organization":"xAI","harness":"","revision":"v1.1","model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"id":"proofbench_external.csv:28","model":"Gemini 3.5 Flash","version":"gemini-3.5-flash_high","effort":"high","score":0.31,"percent":31.0,"cost":0.54137,"seconds":394.5,"provider":"Google","date":"2026-05-19","organization":"Google DeepMind","harness":"","revision":"v1.1","model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"id":"proofbench_external.csv:29","model":"Qwen3.7 Max","version":"qwen3.7-max","effort":null,"score":0.26,"percent":26.0,"cost":2.462069,"seconds":771.408,"provider":"Alibaba","date":"2026-05-19","organization":"Alibaba","harness":"","revision":"v1.1","model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"id":"proofbench_external.csv:31","model":"GLM-5.1","version":"glm-5.1","effort":null,"score":0.22222000000000003,"percent":22.222,"cost":0.584577,"seconds":4755.746,"provider":"Zhipu AI","date":"2026-04-07","organization":"Z.ai (Zhipu AI)","harness":"","revision":"v1.1","model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"id":"proofbench_external.csv:32","model":"MiMo-V2.5-Pro","version":"mimo-v2.5-pro","effort":null,"score":0.22,"percent":22.0,"cost":0.089853,"seconds":1172.459,"provider":"Xiaomi","date":null,"organization":"Xiaomi Corp","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:33","model":"GPT-5.4 mini","version":"gpt-5.4-mini-2026-03-17_xhigh","effort":"xhigh","score":0.21,"percent":21.0,"cost":0.249441,"seconds":1410.671,"provider":"OpenAI","date":"2026-03-17","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"id":"proofbench_external.csv:35","model":"Gemini 3 Pro Preview","version":"gemini-3-pro-preview","effort":"high","score":0.2,"percent":20.0,"cost":1.229076,"seconds":10383.833,"provider":"Google","date":"2025-11-18","organization":"Google DeepMind","harness":"","revision":"v1.1","model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"id":"proofbench_external.csv:36","model":"Claude Sonnet 4.5","version":"claude-sonnet-4-5-20250929_unknown","effort":"unknown thinking","score":0.19,"percent":19.0,"cost":0.65341,"seconds":12300.813,"provider":"Anthropic","date":"2025-09-29","organization":"Anthropic","harness":"","revision":"v1.1","model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"id":"proofbench_external.csv:37","model":"MiniMax-M3","version":"MiniMax-M3","effort":null,"score":0.18,"percent":18.0,"cost":0.419224,"seconds":718.66,"provider":"MiniMax","date":"2026-06-01","organization":"MiniMax","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:38","model":"GPT-5","version":"gpt-5-2025-08-07_high","effort":"high","score":0.18,"percent":18.0,"cost":0.852361,"seconds":4579.325,"provider":"OpenAI","date":"2025-08-07","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"id":"proofbench_external.csv:39","model":"Muse Spark","version":"muse-spark","effort":null,"score":0.17,"percent":17.0,"cost":0.067086,"seconds":478.958,"provider":"Meta","date":"2026-04-08","organization":"Meta AI","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:40","model":"Kimi K2.6","version":"kimi-k2.6","effort":null,"score":0.16,"percent":16.0,"cost":0.295811,"seconds":1386.754,"provider":"Moonshot AI","date":"2026-04-20","organization":"Moonshot","harness":"","revision":"v1.1","model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"id":"proofbench_external.csv:42","model":"DeepSeek v4 Pro","version":"deepseek-v4-pro_max","effort":"max","score":0.16,"percent":16.0,"cost":0.033594,"seconds":419.9,"provider":"DeepSeek","date":"2026-04-24","organization":"DeepSeek","harness":"","revision":"v1.1","model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"id":"proofbench_external.csv:44","model":"Gemini 3 Flash Preview","version":"gemini-3-flash-preview","effort":"high","score":0.15,"percent":15.0,"cost":0.327286,"seconds":1141.548,"provider":"Google","date":"2025-12-17","organization":"Google DeepMind","harness":"","revision":"v1.1","model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"id":"proofbench_external.csv:45","model":"GPT-5.2","version":"gpt-5.2-2025-12-11_xhigh","effort":"xhigh","score":0.15,"percent":15.0,"cost":1.059689,"seconds":1493.817,"provider":"OpenAI","date":"2025-12-11","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"id":"proofbench_external.csv:46","model":"Grok 4.20","version":"grok-4.20-0309-reasoning","effort":null,"score":0.14,"percent":14.000000000000002,"cost":0.31815,"seconds":494.946,"provider":"SpaceXAI","date":"2026-02-17","organization":"xAI","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:47","model":"Gemini 3.5 Flash-Lite","version":"gemini-3.5-flash-lite","effort":"high","score":0.13,"percent":13.0,"cost":0.132788,"seconds":290.614,"provider":"Google","date":"2026-07-21","organization":"Google DeepMind","harness":"","revision":"v1.1","model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"id":"proofbench_external.csv:48","model":"GPT-5 nano","version":"gpt-5-nano-2025-08-07_high","effort":"high","score":0.12,"percent":12.0,"cost":0.049218,"seconds":3729.837,"provider":"OpenAI","date":"2025-08-07","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"id":"proofbench_external.csv:49","model":"Grok 4.3 Beta","version":"grok-4.3_high","effort":"high","score":0.11,"percent":11.0,"cost":0.244122,"seconds":1643.019,"provider":"SpaceXAI","date":"2026-04-17","organization":"xAI","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:50","model":"GPT-5 mini","version":"gpt-5-mini-2025-08-07_high","effort":"high","score":0.09,"percent":9.0,"cost":0.22021,"seconds":3349.817,"provider":"OpenAI","date":"2025-08-07","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"id":"proofbench_external.csv:51","model":"GPT-5.1-Codex-Max","version":"gpt-5.1-codex-max","effort":"high","score":0.09,"percent":9.0,"cost":0.527356,"seconds":3184.213,"provider":"OpenAI","date":"2025-11-19","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.1-codex-max","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex-max.toml"},{"id":"proofbench_external.csv:52","model":"Mistral Medium 3.5","version":"mistral-medium-2604","effort":"high","score":0.09,"percent":9.0,"cost":1.132768,"seconds":479.048,"provider":"Mistral AI","date":"2026-04-28","organization":"Mistral AI","harness":"","revision":"v1.1","model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"id":"proofbench_external.csv:53","model":"DeepSeek-V3.2 (Thinking; Fireworks)","version":"fireworks/deepseek-v3p2","effort":"none","score":0.08,"percent":8.0,"cost":0.11773,"seconds":3981.084,"provider":"Fireworks AI","date":"2025-12-01","organization":"DeepSeek","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:54","model":"GLM-4.7","version":"glm-4.7","effort":null,"score":0.06,"percent":6.0,"cost":0.14404,"seconds":3364.789,"provider":"Zhipu AI","date":"2025-12-22","organization":"Z.ai (Zhipu AI)","harness":"","revision":"v1.1","model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"id":"proofbench_external.csv:55","model":"Inkling Small","version":"inkling-small","effort":"0.99","score":0.06,"percent":6.0,"cost":0.082234,"seconds":507.301,"provider":"Thinkingmachines","date":"2026-07-15","organization":"Thinking Machines","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:56","model":"GPT-5.4 nano","version":"gpt-5.4-nano-2026-03-17_high","effort":"high","score":0.05,"percent":5.0,"cost":0.015894,"seconds":525.432,"provider":"OpenAI","date":"2026-03-17","organization":"OpenAI","harness":"","revision":"v1.1","model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"id":"proofbench_external.csv:57","model":"Grok 4.1 Fast","version":"grok-4-1-fast-reasoning","effort":null,"score":0.04,"percent":4.0,"cost":0.096578,"seconds":1421.041,"provider":"SpaceXAI","date":"2025-11-19","organization":"xAI","harness":"","revision":"v1.1","model_id":"xai/grok-4.1-fast","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.1-fast.toml"},{"id":"proofbench_external.csv:58","model":"MiniMax-M2.5","version":"MiniMax-M2.5","effort":null,"score":0.04,"percent":4.0,"cost":0.057497,"seconds":4514.143,"provider":"MiniMax","date":"2026-02-12","organization":"MiniMax","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:59","model":"MiniMax-M2.7","version":"MiniMax-M2.7","effort":null,"score":0.03,"percent":3.0,"cost":0.04293,"seconds":1634.068,"provider":"MiniMax","date":"2026-03-18","organization":"MiniMax","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:60","model":"Nemotron 3 Ultra","version":"nemotron-3-ultra","effort":null,"score":0.02,"percent":2.0,"cost":null,"seconds":2552.288,"provider":"Nvidia","date":"2026-06-04","organization":"NVIDIA","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:61","model":"Laguna XS.2","version":"laguna-xs.2","effort":null,"score":0.0,"percent":0.0,"cost":0.024998,"seconds":315.721,"provider":"Poolside","date":"2026-04-28","organization":"Poolside","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:62","model":"Laguna M.1","version":"laguna-m.1","effort":null,"score":0.0,"percent":0.0,"cost":0.067403,"seconds":423.12,"provider":"Poolside","date":"2026-04-28","organization":"Poolside","harness":"","revision":"v1.1"},{"id":"proofbench_external.csv:63","model":"Inkling","version":"Inkling","effort":null,"score":0.0,"percent":0.0,"cost":0.198639,"seconds":436.43,"provider":"Thinkingmachines","date":"2026-07-15","organization":"Thinking Machines","harness":"","revision":"v1.1"}],"cost_label":"Cost per test (USD)","time_label":"Latency (seconds)"}],"index":{"id":"epoch-capabilities-index","name":"Epoch Capabilities Index","abbr":"ECI","description":"One overall score per model, worked out from all the benchmarks Epoch runs. Higher is better. The scale has no maximum, so a score only means something next to another score, never on its own.","entrants":266,"rows":[{"model":"GPT-6 Astra","score":166.31,"ci_low":163.0,"ci_high":171.88,"date":"2026-09-03","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":1,"training_compute_flop":null,"frontier_days":12,"lag_months":0.0,"lag_low":0.0,"lag_high":2.8,"frontier_date":"2026-09-03","frontier_model":"GPT-6 Astra","at_frontier":true,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5.1","score":164.47,"ci_low":161.36,"ci_high":168.28,"date":"2026-09-01","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":2,"training_compute_flop":null,"frontier_days":2,"lag_months":0.0,"lag_low":0.0,"lag_high":4.3,"frontier_date":"2026-09-01","frontier_model":"Claude Fable 5.1","at_frontier":true,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Fable 5","score":163.27,"ci_low":160.54,"ci_high":167.1,"date":"2026-06-09","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":3,"training_compute_flop":null,"frontier_days":84,"lag_months":0.0,"lag_low":0.0,"lag_high":1.5,"frontier_date":"2026-06-09","frontier_model":"Claude Fable 5","at_frontier":true,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":162.3,"ci_low":159.9,"ci_high":165.74,"date":"2026-07-24","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":4,"training_compute_flop":null,"frontier_days":0,"lag_months":1.5,"lag_low":1.5,"lag_high":3.0,"frontier_date":"2026-06-09","frontier_model":"Claude Fable 5","at_frontier":false,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.5 Pro","score":162.25,"ci_low":159.32,"ci_high":165.76,"date":"2026-04-23","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":5,"training_compute_flop":null,"frontier_days":47,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2026-04-23","frontier_model":"GPT-5.5 Pro","at_frontier":true,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"GPT-5.6 Sol","score":161.81,"ci_low":159.57,"ci_high":165.2,"date":"2026-07-09","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":6,"training_compute_flop":null,"frontier_days":0,"lag_months":2.5,"lag_low":2.5,"lag_high":2.5,"frontier_date":"2026-04-23","frontier_model":"GPT-5.5 Pro","at_frontier":false,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.6 Terra","score":159.14,"ci_low":157.01,"ci_high":161.84,"date":"2026-07-09","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":7,"training_compute_flop":null,"frontier_days":0,"lag_months":2.5,"lag_low":2.5,"lag_high":4.1,"frontier_date":"2026-04-23","frontier_model":"GPT-5.5 Pro","at_frontier":false,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.5","score":159.12,"ci_low":156.91,"ci_high":161.95,"date":"2026-04-23","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":8,"training_compute_flop":null,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":1.6,"frontier_date":"2026-04-23","frontier_model":"GPT-5.5 Pro","at_frontier":true,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4 Pro","score":158.95,"ci_low":156.73,"ci_high":161.7,"date":"2026-03-05","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":9,"training_compute_flop":null,"frontier_days":49,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":true,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Claude Opus 4.8","score":158.3,"ci_low":156.43,"ci_high":160.71,"date":"2026-05-28","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":10,"training_compute_flop":null,"frontier_days":0,"lag_months":2.8,"lag_low":1.1,"lag_high":3.7,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":false,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Kimi K3","score":157.63,"ci_low":155.38,"ci_high":160.45,"date":"2026-07-16","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["China"],"rank":11,"training_compute_flop":2.0001e+25,"frontier_days":0,"lag_months":4.4,"lag_low":2.8,"lag_high":7.1,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":false,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.7 Flash","score":157.44,"ci_low":155.66,"ci_high":159.78,"date":"2026-08-13","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":12,"training_compute_flop":null,"frontier_days":0,"lag_months":5.3,"lag_low":3.7,"lag_high":6.2,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":false,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GPT-5.4","score":156.86,"ci_low":155.21,"ci_high":159.0,"date":"2026-03-05","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":13,"training_compute_flop":null,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":2.8,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":true,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Qwen 3.8 Max","score":156.62,"ci_low":154.85,"ci_high":158.75,"date":"2026-08-02","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":14,"training_compute_flop":null,"frontier_days":0,"lag_months":4.9,"lag_low":4.9,"lag_high":7.7,"frontier_date":"2026-03-05","frontier_model":"GPT-5.4 Pro","at_frontier":false,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GPT-5.3 Codex","score":156.58,"ci_low":153.91,"ci_high":159.73,"date":"2026-02-05","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":15,"training_compute_flop":null,"frontier_days":28,"lag_months":0.0,"lag_low":0.0,"lag_high":1.8,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":true,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"Gemini 3.8 Flash","score":156.54,"ci_low":154.58,"ci_high":161.97,"date":"2026-09-02","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":16,"training_compute_flop":null,"frontier_days":0,"lag_months":6.9,"lag_low":4.3,"lag_high":8.7,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Grok 4.6","score":156.35,"ci_low":154.77,"ci_high":158.88,"date":"2026-08-12","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":17,"training_compute_flop":null,"frontier_days":0,"lag_months":6.2,"lag_low":5.3,"lag_high":8.0,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Claude Opus 4.7","score":156.34,"ci_low":154.57,"ci_high":158.5,"date":"2026-04-16","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":18,"training_compute_flop":null,"frontier_days":0,"lag_months":2.3,"lag_low":1.4,"lag_high":4.1,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.6 Luna","score":156.31,"ci_low":154.06,"ci_high":158.86,"date":"2026-07-09","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":19,"training_compute_flop":null,"frontier_days":0,"lag_months":5.1,"lag_low":4.1,"lag_high":6.9,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Sonnet 5","score":156.22,"ci_low":153.63,"ci_high":158.63,"date":"2026-06-30","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":20,"training_compute_flop":null,"frontier_days":0,"lag_months":4.8,"lag_low":3.8,"lag_high":6.6,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Muse Spark 1.2","score":155.48,"ci_low":153.9,"ci_high":158.64,"date":"2026-08-05","organization":"Meta AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":21,"training_compute_flop":null,"frontier_days":0,"lag_months":5.9,"lag_low":5.0,"lag_high":7.8,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"DeepSeek V4 Pro 0813","score":155.47,"ci_low":153.78,"ci_high":157.83,"date":"2026-08-13","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":22,"training_compute_flop":9.702e+24,"frontier_days":0,"lag_months":6.2,"lag_low":5.3,"lag_high":8.0,"frontier_date":"2026-02-05","frontier_model":"GPT-5.3 Codex","at_frontier":false,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"GPT-5.2 Pro","score":155.38,"ci_low":153.18,"ci_high":158.22,"date":"2025-12-11","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":23,"training_compute_flop":null,"frontier_days":56,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":true,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"Claude Opus 4.6","score":155.34,"ci_low":153.7,"ci_high":157.21,"date":"2026-02-05","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":24,"training_compute_flop":null,"frontier_days":0,"lag_months":1.8,"lag_low":1.8,"lag_high":1.8,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GLM-5.3","score":155.25,"ci_low":153.18,"ci_high":157.53,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":25,"training_compute_flop":null,"frontier_days":0,"lag_months":8.1,"lag_low":5.3,"lag_high":8.1,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Qwen3.8 Max (0902)","score":155.22,"ci_low":153.57,"ci_high":157.34,"date":"2026-09-01","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":26,"training_compute_flop":null,"frontier_days":0,"lag_months":8.7,"lag_low":5.9,"lag_high":8.7,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false},{"model":"Gemini 3.1 Pro","score":155.0,"ci_low":152.72,"ci_high":157.6,"date":"2026-02-19","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":27,"training_compute_flop":null,"frontier_days":0,"lag_months":2.3,"lag_low":2.3,"lag_high":3.1,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false},{"model":"Gemini 3.5 Flash","score":154.69,"ci_low":152.78,"ci_high":157.03,"date":"2026-05-19","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":28,"training_compute_flop":null,"frontier_days":0,"lag_months":5.2,"lag_low":2.5,"lag_high":6.0,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Muse Spark 1.1","score":154.62,"ci_low":152.58,"ci_high":157.88,"date":"2026-07-09","organization":"Meta AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":29,"training_compute_flop":null,"frontier_days":0,"lag_months":6.9,"lag_low":4.1,"lag_high":7.7,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"DeepSeek V4 Flash 0731","score":154.49,"ci_low":152.4,"ci_high":156.51,"date":"2026-07-31","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":30,"training_compute_flop":2.496e+24,"frontier_days":0,"lag_months":7.6,"lag_low":5.8,"lag_high":8.4,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Gemini 3.6 Flash","score":154.33,"ci_low":152.86,"ci_high":156.24,"date":"2026-07-21","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":31,"training_compute_flop":null,"frontier_days":0,"lag_months":7.3,"lag_low":5.5,"lag_high":8.0,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Grok 4.5","score":153.92,"ci_low":152.49,"ci_high":156.04,"date":"2026-07-08","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":32,"training_compute_flop":null,"frontier_days":0,"lag_months":6.9,"lag_low":5.0,"lag_high":7.6,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Qwen3.7-Max","score":153.73,"ci_low":151.88,"ci_high":155.88,"date":"2026-05-19","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":33,"training_compute_flop":null,"frontier_days":0,"lag_months":5.2,"lag_low":3.4,"lag_high":6.0,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":false,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"GPT-5.2","score":153.51,"ci_low":151.75,"ci_high":155.43,"date":"2025-12-11","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":34,"training_compute_flop":null,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.8,"frontier_date":"2025-12-11","frontier_model":"GPT-5.2 Pro","at_frontier":true,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3 Pro","score":153.0,"ci_low":151.05,"ci_high":155.08,"date":"2025-11-18","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":35,"training_compute_flop":null,"frontier_days":23,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":true},{"model":"Claude Sonnet 4.6","score":152.25,"ci_low":149.69,"ci_high":154.47,"date":"2026-02-17","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":36,"training_compute_flop":null,"frontier_days":0,"lag_months":3.0,"lag_low":2.2,"lag_high":6.4,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Muse Spark","score":152.12,"ci_low":150.27,"ci_high":155.59,"date":"2026-04-08","organization":"Meta AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":37,"training_compute_flop":null,"frontier_days":0,"lag_months":4.6,"lag_low":2.0,"lag_high":6.0,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false},{"model":"Grok 4.20","score":152.03,"ci_low":149.21,"ci_high":154.36,"date":"2026-02-17","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":38,"training_compute_flop":null,"frontier_days":0,"lag_months":3.0,"lag_low":2.2,"lag_high":6.4,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false},{"model":"GLM-5.2","score":151.86,"ci_low":150.21,"ci_high":154.16,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":39,"training_compute_flop":null,"frontier_days":0,"lag_months":6.9,"lag_low":6.1,"lag_high":8.3,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Gemini 3 Flash","score":151.84,"ci_low":150.57,"ci_high":153.61,"date":"2025-12-17","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":40,"training_compute_flop":null,"frontier_days":0,"lag_months":1.0,"lag_low":0.2,"lag_high":1.0,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false},{"model":"GLM-5.3-Flash","score":151.45,"ci_low":149.28,"ci_high":153.78,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":41,"training_compute_flop":3.24e+24,"frontier_days":0,"lag_months":9.0,"lag_low":8.3,"lag_high":12.4,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Kimi K2.6","score":151.0,"ci_low":149.14,"ci_high":152.7,"date":"2026-04-20","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":42,"training_compute_flop":null,"frontier_days":0,"lag_months":5.0,"lag_low":5.0,"lag_high":8.4,"frontier_date":"2025-11-18","frontier_model":"Gemini 3 Pro","at_frontier":false,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"GPT-5 Pro","score":150.29,"ci_low":148.83,"ci_high":152.9,"date":"2025-10-07","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":43,"training_compute_flop":null,"frontier_days":42,"lag_months":0.0,"lag_low":0.0,"lag_high":2.0,"frontier_date":"2025-10-07","frontier_model":"GPT-5 Pro","at_frontier":true,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"Inkling-Small","score":150.15,"ci_low":147.34,"ci_high":152.09,"date":"2026-07-15","organization":"Thinking Machines","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":44,"training_compute_flop":null,"frontier_days":0,"lag_months":9.2,"lag_low":7.9,"lag_high":13.1,"frontier_date":"2025-10-07","frontier_model":"GPT-5 Pro","at_frontier":false},{"model":"Claude Opus 4.5","score":150.11,"ci_low":148.12,"ci_high":152.85,"date":"2025-11-24","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":45,"training_compute_flop":null,"frontier_days":0,"lag_months":1.6,"lag_low":0.2,"lag_high":3.6,"frontier_date":"2025-10-07","frontier_model":"GPT-5 Pro","at_frontier":false,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Kimi K2.7 Code","score":150.07,"ci_low":148.68,"ci_high":151.93,"date":"2026-06-12","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":46,"training_compute_flop":null,"frontier_days":0,"lag_months":8.1,"lag_low":6.8,"lag_high":10.2,"frontier_date":"2025-10-07","frontier_model":"GPT-5 Pro","at_frontier":false,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5","score":150.0,"ci_low":null,"ci_high":null,"date":"2025-08-07","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":47,"training_compute_flop":6.6e+25,"frontier_days":61,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":true,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GLM-5.1","score":149.71,"ci_low":147.97,"ci_high":151.41,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":48,"training_compute_flop":null,"frontier_days":0,"lag_months":8.0,"lag_low":4.6,"lag_high":8.0,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"GPT-5.1","score":149.64,"ci_low":148.45,"ci_high":151.03,"date":"2025-11-13","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":49,"training_compute_flop":null,"frontier_days":0,"lag_months":3.2,"lag_low":3.2,"lag_high":3.2,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Qwen 3.6 Max (Preview)","score":149.27,"ci_low":147.69,"ci_high":152.44,"date":"2026-04-20","organization":"Alibaba","open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":["China"],"rank":50,"training_compute_flop":null,"frontier_days":0,"lag_months":8.4,"lag_low":5.0,"lag_high":8.4,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false},{"model":"Grok 4.3 Beta","score":149.17,"ci_low":147.78,"ci_high":150.85,"date":"2026-04-17","organization":"xAI","open_weights":false,"accessibility":"Hosted access (no API)","accessibility_group":"Closed weights","countries":["United States of America"],"rank":51,"training_compute_flop":null,"frontier_days":0,"lag_months":8.3,"lag_low":4.9,"lag_high":8.3,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false},{"model":"DeepSeek-V4-Pro","score":149.09,"ci_low":147.52,"ci_high":150.74,"date":"2026-04-24","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":52,"training_compute_flop":9.702e+24,"frontier_days":0,"lag_months":8.5,"lag_low":5.2,"lag_high":8.5,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5.4 Mini","score":148.96,"ci_low":147.33,"ci_high":150.68,"date":"2026-03-17","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":53,"training_compute_flop":null,"frontier_days":0,"lag_months":7.3,"lag_low":3.9,"lag_high":9.2,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Inkling","score":148.64,"ci_low":145.95,"ci_high":150.6,"date":"2026-07-15","organization":"Thinking Machines","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":54,"training_compute_flop":1.845e+24,"frontier_days":0,"lag_months":11.2,"lag_low":7.9,"lag_high":14.9,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false},{"model":"Kimi K2.5","score":148.01,"ci_low":146.58,"ci_high":149.35,"date":"2026-01-27","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":55,"training_compute_flop":5.8e+24,"frontier_days":0,"lag_months":5.7,"lag_low":5.7,"lag_high":9.4,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false},{"model":"Qwen 3.6 Plus","score":147.67,"ci_low":145.74,"ci_high":149.56,"date":"2026-03-31","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":56,"training_compute_flop":null,"frontier_days":0,"lag_months":7.8,"lag_low":7.8,"lag_high":11.5,"frontier_date":"2025-08-07","frontier_model":"GPT-5","at_frontier":false,"model_id":"alibaba/qwen3.6-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-plus.toml"},{"model":"o3-pro","score":147.45,"ci_low":145.85,"ci_high":149.7,"date":"2025-06-10","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":57,"training_compute_flop":null,"frontier_days":58,"lag_months":0.0,"lag_low":0.0,"lag_high":1.8,"frontier_date":"2025-06-10","frontier_model":"o3-pro","at_frontier":true,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"Qwen3.7-Plus","score":147.41,"ci_low":145.92,"ci_high":148.83,"date":"2026-06-02","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":58,"training_compute_flop":null,"frontier_days":0,"lag_months":11.7,"lag_low":9.8,"lag_high":13.5,"frontier_date":"2025-06-10","frontier_model":"o3-pro","at_frontier":false,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Qwen3.5 397B-A17B","score":146.96,"ci_low":145.43,"ci_high":148.62,"date":"2026-02-13","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":59,"training_compute_flop":null,"frontier_days":0,"lag_months":8.1,"lag_low":6.2,"lag_high":10.0,"frontier_date":"2025-06-10","frontier_model":"o3-pro","at_frontier":false,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"o3","score":146.91,"ci_low":145.23,"ci_high":148.66,"date":"2025-04-16","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":60,"training_compute_flop":null,"frontier_days":55,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":true,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude Sonnet 4.5","score":146.84,"ci_low":145.46,"ci_high":148.5,"date":"2025-09-29","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":61,"training_compute_flop":null,"frontier_days":0,"lag_months":5.5,"lag_low":1.7,"lag_high":5.5,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Qwen 3.5 Plus (hosted 397B-A17B)","score":146.73,"ci_low":144.76,"ci_high":148.09,"date":"2026-02-16","organization":"Alibaba","open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":["China"],"rank":62,"training_compute_flop":null,"frontier_days":0,"lag_months":10.1,"lag_low":6.3,"lag_high":10.1,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"MiniMax-M2.5","score":146.51,"ci_low":138.67,"ci_high":147.85,"date":"2026-02-12","organization":"MiniMax","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":63,"training_compute_flop":null,"frontier_days":0,"lag_months":9.9,"lag_low":6.2,"lag_high":13.9,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Qwen3.6 27B","score":146.49,"ci_low":144.26,"ci_high":147.95,"date":"2026-04-22","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":64,"training_compute_flop":null,"frontier_days":0,"lag_months":12.2,"lag_low":8.5,"lag_high":12.2,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"MiniMax-M3","score":146.47,"ci_low":142.57,"ci_high":149.55,"date":"2026-06-01","organization":"MiniMax","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":65,"training_compute_flop":null,"frontier_days":0,"lag_months":13.5,"lag_low":9.8,"lag_high":14.0,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Grok 4","score":146.45,"ci_low":144.62,"ci_high":148.2,"date":"2025-07-09","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":66,"training_compute_flop":5.0000000000001e+26,"frontier_days":0,"lag_months":2.8,"lag_low":2.8,"lag_high":2.8,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Nemotron 3 Ultra","score":146.27,"ci_low":144.01,"ci_high":148.3,"date":"2026-06-04","organization":"NVIDIA","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":67,"training_compute_flop":6.6e+24,"frontier_days":0,"lag_months":13.6,"lag_low":9.9,"lag_high":14.1,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"DeepSeek-V3.2","score":146.19,"ci_low":144.45,"ci_high":147.44,"date":"2025-12-01","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":68,"training_compute_flop":4.2e+24,"frontier_days":0,"lag_months":7.5,"lag_low":5.7,"lag_high":7.5,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"model":"DeepSeek-V4-Flash","score":146.11,"ci_low":144.22,"ci_high":148.12,"date":"2026-04-24","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":69,"training_compute_flop":2.496e+24,"frontier_days":0,"lag_months":12.3,"lag_low":8.5,"lag_high":12.3,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"model":"GLM-5","score":145.85,"ci_low":144.17,"ci_high":147.76,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":70,"training_compute_flop":6.84e+24,"frontier_days":0,"lag_months":9.9,"lag_low":6.2,"lag_high":9.9,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5.4 Nano","score":145.85,"ci_low":143.7,"ci_high":147.77,"date":"2026-03-17","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":71,"training_compute_flop":null,"frontier_days":0,"lag_months":11.0,"lag_low":7.3,"lag_high":11.5,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"MiniMax-M2.7","score":145.8,"ci_low":138.57,"ci_high":147.75,"date":"2026-03-18","organization":"MiniMax","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["China"],"rank":72,"training_compute_flop":null,"frontier_days":0,"lag_months":11.0,"lag_low":7.3,"lag_high":15.0,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Kimi K2 Thinking","score":145.76,"ci_low":143.25,"ci_high":147.14,"date":"2025-11-06","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":73,"training_compute_flop":4.2e+24,"frontier_days":0,"lag_months":6.7,"lag_low":4.9,"lag_high":7.2,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"o4-mini","score":145.65,"ci_low":143.36,"ci_high":147.36,"date":"2025-04-16","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":74,"training_compute_flop":null,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.5,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":true,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"GPT-5 mini","score":145.52,"ci_low":143.6,"ci_high":147.15,"date":"2025-08-07","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":75,"training_compute_flop":null,"frontier_days":0,"lag_months":3.7,"lag_low":1.9,"lag_high":4.2,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":145.26,"ci_low":143.82,"ci_high":146.79,"date":"2025-06-05","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":76,"training_compute_flop":null,"frontier_days":0,"lag_months":1.6,"lag_low":1.6,"lag_high":2.2,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Gemini 3.5 Flash-Lite","score":145.13,"ci_low":142.76,"ci_high":146.76,"date":"2026-07-21","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":77,"training_compute_flop":null,"frontier_days":0,"lag_months":15.1,"lag_low":15.1,"lag_high":15.7,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"DeepSeek-V3.2-Exp","score":145.02,"ci_low":142.61,"ci_high":147.24,"date":"2025-09-29","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":78,"training_compute_flop":4.18e+24,"frontier_days":0,"lag_months":5.5,"lag_low":3.6,"lag_high":6.0,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Qwen3.7 Flash","score":144.63,"ci_low":142.86,"ci_high":147.57,"date":"2026-07-27","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":79,"training_compute_flop":null,"frontier_days":0,"lag_months":15.3,"lag_low":11.6,"lag_high":15.9,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"Gemini 3.1 Flash-Lite","score":144.49,"ci_low":142.78,"ci_high":146.17,"date":"2026-03-03","organization":"Google","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":80,"training_compute_flop":null,"frontier_days":0,"lag_months":10.5,"lag_low":10.5,"lag_high":11.1,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Grok 4 Fast","score":144.21,"ci_low":142.22,"ci_high":146.36,"date":"2025-09-19","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":81,"training_compute_flop":null,"frontier_days":0,"lag_months":5.1,"lag_low":5.1,"lag_high":5.7,"frontier_date":"2025-04-16","frontier_model":"o3","at_frontier":false},{"model":"Gemini 2.5 Pro (Mar 2025)","score":144.16,"ci_low":142.19,"ci_high":147.04,"date":"2025-03-31","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":82,"training_compute_flop":null,"frontier_days":16,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":true},{"model":"Claude Opus 4.1","score":144.11,"ci_low":142.06,"ci_high":145.99,"date":"2025-08-05","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":83,"training_compute_flop":null,"frontier_days":0,"lag_months":4.2,"lag_low":3.6,"lag_high":4.2,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Qwen 3.5 Flash (hosted 35B-A3B)","score":144.0,"ci_low":142.32,"ci_high":145.75,"date":"2026-02-25","organization":"Alibaba","open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":["China"],"rank":84,"training_compute_flop":null,"frontier_days":0,"lag_months":10.9,"lag_low":10.3,"lag_high":10.9,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":143.88,"ci_low":142.1,"ci_high":145.38,"date":"2025-07-25","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":85,"training_compute_flop":4.752e+24,"frontier_days":0,"lag_months":3.8,"lag_low":3.3,"lag_high":3.8,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false},{"model":"Qwen 3.6 35B-A3B","score":143.86,"ci_low":141.79,"ci_high":146.13,"date":"2026-04-14","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":86,"training_compute_flop":null,"frontier_days":0,"lag_months":12.5,"lag_low":11.9,"lag_high":15.9,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"GLM-4.7","score":143.46,"ci_low":141.57,"ci_high":145.43,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":87,"training_compute_flop":4.42e+24,"frontier_days":0,"lag_months":8.7,"lag_low":8.2,"lag_high":12.2,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"Qwen 3.6 Flash","score":143.27,"ci_low":140.98,"ci_high":144.97,"date":"2026-04-27","organization":"Alibaba","open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":["China"],"rank":88,"training_compute_flop":null,"frontier_days":0,"lag_months":12.9,"lag_low":12.4,"lag_high":16.3,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"alibaba/qwen3.6-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-flash.toml"},{"model":"Gemini 2.5 Flash (Sep 2025)","score":142.99,"ci_low":139.67,"ci_high":146.63,"date":"2025-09-25","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":89,"training_compute_flop":null,"frontier_days":0,"lag_months":5.8,"lag_low":5.3,"lag_high":9.3,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false},{"model":"Claude Opus 4","score":142.68,"ci_low":140.74,"ci_high":144.27,"date":"2025-05-22","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":90,"training_compute_flop":null,"frontier_days":0,"lag_months":1.7,"lag_low":1.2,"lag_high":5.1,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Gemma 4 31B IT","score":142.66,"ci_low":140.48,"ci_high":145.02,"date":"2026-04-02","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":91,"training_compute_flop":null,"frontier_days":0,"lag_months":12.1,"lag_low":11.5,"lag_high":15.5,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"Qwen3.5-35B-A3B","score":142.53,"ci_low":140.39,"ci_high":144.77,"date":"2026-02-24","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":92,"training_compute_flop":null,"frontier_days":0,"lag_months":10.8,"lag_low":10.3,"lag_high":14.3,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"GPT-5.5 Instant","score":142.5,"ci_low":140.25,"ci_high":145.35,"date":"2026-05-05","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":93,"training_compute_flop":null,"frontier_days":0,"lag_months":13.1,"lag_low":12.6,"lag_high":16.6,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"Gemini 2.5 Pro (May 2025)","score":142.47,"ci_low":139.42,"ci_high":145.42,"date":"2025-05-06","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":94,"training_compute_flop":null,"frontier_days":0,"lag_months":1.2,"lag_low":0.7,"lag_high":4.6,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false},{"model":"Qwen3-Max","score":142.43,"ci_low":140.4,"ci_high":144.98,"date":"2025-09-24","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":95,"training_compute_flop":1.512e+25,"frontier_days":0,"lag_months":5.8,"lag_low":5.3,"lag_high":9.2,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"alibaba/qwen3-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-max.toml"},{"model":"Claude Haiku 4.5","score":142.37,"ci_low":139.8,"ci_high":144.07,"date":"2025-10-15","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":96,"training_compute_flop":null,"frontier_days":0,"lag_months":6.5,"lag_low":6.5,"lag_high":9.9,"frontier_date":"2025-03-31","frontier_model":"Gemini 2.5 Pro (Mar 2025)","at_frontier":false,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"o1","score":141.86,"ci_low":140.15,"ci_high":143.11,"date":"2024-12-17","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":97,"training_compute_flop":null,"frontier_days":104,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":true,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Gemma 4 26B A4B","score":141.86,"ci_low":138.61,"ci_high":143.65,"date":"2026-04-02","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":98,"training_compute_flop":null,"frontier_days":0,"lag_months":15.5,"lag_low":12.1,"lag_high":15.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Claude Sonnet 4","score":141.69,"ci_low":139.18,"ci_high":143.08,"date":"2025-05-22","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":99,"training_compute_flop":null,"frontier_days":0,"lag_months":5.1,"lag_low":1.7,"lag_high":5.1,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":141.54,"ci_low":139.66,"ci_high":143.04,"date":"2025-05-20","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":100,"training_compute_flop":null,"frontier_days":0,"lag_months":5.1,"lag_low":1.6,"lag_high":5.1,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"DeepSeek-R1 (May 2025)","score":141.29,"ci_low":139.31,"ci_high":142.74,"date":"2025-05-28","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":101,"training_compute_flop":4.020010000000001e+24,"frontier_days":0,"lag_months":5.3,"lag_low":1.9,"lag_high":5.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Mistral Medium 3.5","score":141.28,"ci_low":139.38,"ci_high":143.28,"date":"2026-04-28","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["France"],"rank":102,"training_compute_flop":null,"frontier_days":0,"lag_months":16.3,"lag_low":12.9,"lag_high":16.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"Claude 3.7 Sonnet","score":141.16,"ci_low":139.15,"ci_high":142.7,"date":"2025-02-24","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":103,"training_compute_flop":3.35e+25,"frontier_days":0,"lag_months":2.3,"lag_low":2.3,"lag_high":2.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"GLM-4.6","score":140.78,"ci_low":135.2,"ci_high":142.57,"date":"2025-09-30","organization":"Z.ai (Zhipu AI),Tsinghua University","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":104,"training_compute_flop":4.42e+24,"frontier_days":0,"lag_months":9.4,"lag_low":6.0,"lag_high":12.6,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"zhipuai/glm-4.6","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.6.toml"},{"model":"Gemini 2.5 Flash (Jun 2025)","score":140.53,"ci_low":138.48,"ci_high":141.94,"date":"2025-06-17","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":105,"training_compute_flop":null,"frontier_days":0,"lag_months":6.0,"lag_low":2.6,"lag_high":6.0,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"o3-mini","score":140.35,"ci_low":137.21,"ci_high":141.89,"date":"2025-01-31","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":106,"training_compute_flop":null,"frontier_days":0,"lag_months":1.5,"lag_low":1.5,"lag_high":1.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Grok-3 mini","score":140.35,"ci_low":138.05,"ci_high":142.03,"date":"2025-06-24","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":107,"training_compute_flop":null,"frontier_days":0,"lag_months":6.2,"lag_low":2.8,"lag_high":6.2,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Kimi K2 (Jul 2025)","score":140.11,"ci_low":137.03,"ci_high":141.99,"date":"2025-07-12","organization":"Moonshot","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":108,"training_compute_flop":2.976e+24,"frontier_days":0,"lag_months":6.8,"lag_low":3.4,"lag_high":6.8,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"gpt-oss-120b","score":140.1,"ci_low":136.28,"ci_high":142.7,"date":"2025-08-05","organization":"OpenAI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":109,"training_compute_flop":4.94e+24,"frontier_days":0,"lag_months":7.6,"lag_low":4.2,"lag_high":7.6,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Gemini 2.5 Flash (Apr 2025)","score":139.97,"ci_low":135.94,"ci_high":142.06,"date":"2025-04-17","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":110,"training_compute_flop":null,"frontier_days":0,"lag_months":4.0,"lag_low":0.6,"lag_high":4.0,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"DeepSeek-V3.1","score":139.92,"ci_low":136.28,"ci_high":143.28,"date":"2025-08-21","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":111,"training_compute_flop":3.594058e+24,"frontier_days":0,"lag_months":8.1,"lag_low":4.7,"lag_high":8.1,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","score":139.64,"ci_low":135.51,"ci_high":141.17,"date":"2025-07-30","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":112,"training_compute_flop":6.48e+23,"frontier_days":0,"lag_months":7.4,"lag_low":7.4,"lag_high":10.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Qwen3.5-9B","score":139.44,"ci_low":136.67,"ci_high":141.53,"date":"2026-02-24","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":113,"training_compute_flop":null,"frontier_days":0,"lag_months":14.3,"lag_low":14.3,"lag_high":14.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"GPT-5 nano","score":139.39,"ci_low":135.2,"ci_high":141.56,"date":"2025-08-07","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":114,"training_compute_flop":null,"frontier_days":0,"lag_months":7.7,"lag_low":7.7,"lag_high":10.8,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Qwen3-235B-A22B","score":139.35,"ci_low":136.34,"ci_high":140.95,"date":"2025-04-28","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":115,"training_compute_flop":4.752e+24,"frontier_days":0,"lag_months":4.3,"lag_low":4.3,"lag_high":4.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"DeepSeek-R1","score":138.97,"ci_low":136.9,"ci_high":140.47,"date":"2025-01-20","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":116,"training_compute_flop":3.5e+24,"frontier_days":0,"lag_months":1.1,"lag_low":1.1,"lag_high":1.1,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":138.92,"ci_low":135.97,"ci_high":140.87,"date":"2025-07-25","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":117,"training_compute_flop":4.752e+24,"frontier_days":0,"lag_months":7.2,"lag_low":7.2,"lag_high":7.2,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Qwen3-32B","score":138.51,"ci_low":135.04,"ci_high":140.45,"date":"2025-04-29","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":118,"training_compute_flop":7.0848e+24,"frontier_days":0,"lag_months":4.4,"lag_low":4.4,"lag_high":7.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Grok 3","score":138.3,"ci_low":135.93,"ci_high":139.83,"date":"2025-04-09","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":119,"training_compute_flop":3.5e+26,"frontier_days":0,"lag_months":3.7,"lag_low":3.7,"lag_high":3.7,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Qwen3-14B","score":138.24,"ci_low":133.78,"ci_high":139.91,"date":"2025-04-29","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":120,"training_compute_flop":3.1968e+24,"frontier_days":0,"lag_months":4.4,"lag_low":4.4,"lag_high":7.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"gpt-oss-20b","score":137.8,"ci_low":133.26,"ci_high":139.73,"date":"2025-08-05","organization":"OpenAI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":121,"training_compute_flop":5.49e+23,"frontier_days":0,"lag_months":7.6,"lag_low":7.6,"lag_high":10.7,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"},{"model":"QwQ-32B","score":137.6,"ci_low":132.9,"ci_high":141.53,"date":"2025-03-05","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":122,"training_compute_flop":3.51e+24,"frontier_days":0,"lag_months":2.6,"lag_low":2.6,"lag_high":5.7,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"model":"DeepSeek-R1-Distill-Qwen-32B","score":137.42,"ci_low":131.53,"ci_high":138.84,"date":"2025-01-20","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":123,"training_compute_flop":null,"frontier_days":0,"lag_months":1.1,"lag_low":1.1,"lag_high":4.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"deepseek/deepseek-r1-distill-qwen-32b","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1-distill-qwen-32b.toml"},{"model":"Qwen3-30B-A3B-Instruct (Jul 2025)","score":137.42,"ci_low":131.98,"ci_high":139.5,"date":"2025-07-29","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":124,"training_compute_flop":6.48e+23,"frontier_days":0,"lag_months":7.4,"lag_low":7.4,"lag_high":10.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"GPT-4.1","score":136.8,"ci_low":134.37,"ci_high":138.51,"date":"2025-04-14","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":125,"training_compute_flop":null,"frontier_days":0,"lag_months":3.9,"lag_low":3.9,"lag_high":7.0,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"GPT-4.5","score":136.75,"ci_low":134.26,"ci_high":138.84,"date":"2025-02-27","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":126,"training_compute_flop":3.8e+26,"frontier_days":0,"lag_months":2.4,"lag_low":2.4,"lag_high":5.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"Qwen3-30B-A3B","score":136.19,"ci_low":130.67,"ci_high":138.63,"date":"2025-04-29","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":127,"training_compute_flop":6.48e+23,"frontier_days":0,"lag_months":4.4,"lag_low":4.4,"lag_high":7.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false,"model_id":"alibaba/qwen3-30b-a3b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-30b-a3b.toml"},{"model":"Qwen3-8B","score":136.18,"ci_low":130.64,"ci_high":138.29,"date":"2025-04-28","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":128,"training_compute_flop":1.7712e+24,"frontier_days":0,"lag_months":4.3,"lag_low":4.3,"lag_high":7.5,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"DeepSeek-V3 (Mar 2025)","score":135.95,"ci_low":133.32,"ci_high":137.9,"date":"2025-03-24","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":129,"training_compute_flop":3.3e+24,"frontier_days":0,"lag_months":3.2,"lag_low":3.2,"lag_high":6.3,"frontier_date":"2024-12-17","frontier_model":"o1","at_frontier":false},{"model":"o1-mini","score":135.82,"ci_low":132.6,"ci_high":137.28,"date":"2024-09-12","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":130,"training_compute_flop":null,"frontier_days":96,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":true},{"model":"DeepSeek-R1-Distill-Qwen-14B","score":135.43,"ci_low":127.32,"ci_high":138.15,"date":"2025-01-20","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":131,"training_compute_flop":null,"frontier_days":0,"lag_months":4.3,"lag_low":1.1,"lag_high":8.3,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Gemini 2.0 Flash Thinking (Jan 2025)","score":135.37,"ci_low":130.54,"ci_high":138.06,"date":"2025-01-21","organization":"Google DeepMind,Google","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":132,"training_compute_flop":null,"frontier_days":0,"lag_months":4.3,"lag_low":1.1,"lag_high":4.3,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Gemini 2.0 Pro","score":135.06,"ci_low":132.0,"ci_high":136.72,"date":"2025-02-05","organization":"Google DeepMind","open_weights":false,"accessibility":"Hosted access (no API)","accessibility_group":"Closed weights","countries":["United States of America"],"rank":133,"training_compute_flop":null,"frontier_days":0,"lag_months":4.8,"lag_low":1.6,"lag_high":4.8,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"GPT-4.1 mini","score":135.02,"ci_low":131.78,"ci_high":136.65,"date":"2025-04-14","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":134,"training_compute_flop":null,"frontier_days":0,"lag_months":7.0,"lag_low":3.9,"lag_high":7.0,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"o1-preview","score":134.78,"ci_low":131.38,"ci_high":138.44,"date":"2024-09-12","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":135,"training_compute_flop":null,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":true},{"model":"Gemini 2.0 Flash (Dec 2024)","score":134.71,"ci_low":124.75,"ci_high":136.67,"date":"2024-12-11","organization":"Google DeepMind,Google","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":136,"training_compute_flop":null,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":21.0,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Gemini 2.0 Flash (Feb 2025)","score":134.69,"ci_low":132.02,"ci_high":136.71,"date":"2025-02-05","organization":"Google DeepMind,Google","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":137,"training_compute_flop":null,"frontier_days":0,"lag_months":4.8,"lag_low":1.6,"lag_high":4.8,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Mistral Medium 3","score":134.07,"ci_low":130.83,"ci_high":135.56,"date":"2025-05-07","organization":"Mistral AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["France"],"rank":138,"training_compute_flop":null,"frontier_days":0,"lag_months":7.8,"lag_low":7.8,"lag_high":7.8,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false,"model_id":"mistral/mistral-medium-2505","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2505.toml"},{"model":"Gemini 2.5 Flash-Lite (Jun 2025)","score":133.93,"ci_low":130.93,"ci_high":136.34,"date":"2025-06-17","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":139,"training_compute_flop":null,"frontier_days":0,"lag_months":9.1,"lag_low":6.0,"lag_high":9.1,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Claude 3.5 Sonnet (October 2024)","score":133.55,"ci_low":129.7,"ci_high":138.29,"date":"2024-10-22","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":140,"training_compute_flop":null,"frontier_days":0,"lag_months":1.3,"lag_low":1.3,"lag_high":4.1,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Magistral Small 1.0","score":133.19,"ci_low":130.02,"ci_high":134.49,"date":"2025-06-10","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":141,"training_compute_flop":null,"frontier_days":0,"lag_months":8.9,"lag_low":8.9,"lag_high":8.9,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Qwen2.5-Max","score":132.53,"ci_low":129.09,"ci_high":135.74,"date":"2025-01-25","organization":"Alibaba","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["China"],"rank":142,"training_compute_flop":null,"frontier_days":0,"lag_months":4.4,"lag_low":4.4,"lag_high":7.2,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"DeepSeek-V3","score":132.35,"ci_low":128.55,"ci_high":135.57,"date":"2024-12-26","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":143,"training_compute_flop":3.3e+24,"frontier_days":0,"lag_months":3.4,"lag_low":3.4,"lag_high":7.5,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Llama 4 Maverick","score":132.2,"ci_low":128.7,"ci_high":134.05,"date":"2025-04-06","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":144,"training_compute_flop":2.244000000001e+24,"frontier_days":0,"lag_months":6.8,"lag_low":6.8,"lag_high":10.8,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Mistral Small 3.2","score":131.74,"ci_low":126.94,"ci_high":133.88,"date":"2025-06-20","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":145,"training_compute_flop":null,"frontier_days":0,"lag_months":9.2,"lag_low":9.2,"lag_high":14.4,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false,"model_id":"mistral/mistral-small-2506","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-small-2506.toml"},{"model":"Gemini 1.5 Pro (Sept 2024)","score":131.73,"ci_low":128.47,"ci_high":133.31,"date":"2024-09-24","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":146,"training_compute_flop":null,"frontier_days":0,"lag_months":0.4,"lag_low":0.4,"lag_high":4.4,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":true},{"model":"Magistral Small 1.2","score":131.41,"ci_low":127.11,"ci_high":133.67,"date":"2025-09-18","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":147,"training_compute_flop":null,"frontier_days":0,"lag_months":12.2,"lag_low":12.2,"lag_high":17.3,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false,"model_id":"mistral/magistral-small-2509","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/magistral-small-2509.toml"},{"model":"Grok-2 (Dec 2024)","score":130.48,"ci_low":126.21,"ci_high":132.43,"date":"2024-12-12","organization":"xAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":148,"training_compute_flop":2.96e+25,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":10.6,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Phi-4","score":130.42,"ci_low":125.56,"ci_high":132.22,"date":"2024-12-12","organization":"Microsoft Research","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":149,"training_compute_flop":9.3202015e+23,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":21.0,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Gemma 3 27B","score":130.03,"ci_low":125.11,"ci_high":132.2,"date":"2025-03-12","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":150,"training_compute_flop":2.268e+24,"frontier_days":0,"lag_months":5.9,"lag_low":5.9,"lag_high":23.9,"frontier_date":"2024-09-12","frontier_model":"o1-mini","at_frontier":false},{"model":"Claude 3.5 Sonnet","score":130.0,"ci_low":null,"ci_high":null,"date":"2024-06-20","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":151,"training_compute_flop":2.700000000000001e+25,"frontier_days":84,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2024-06-20","frontier_model":"Claude 3.5 Sonnet","at_frontier":true},{"model":"Llama 4 Scout","score":129.64,"ci_low":125.22,"ci_high":131.37,"date":"2025-04-05","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":152,"training_compute_flop":4.08e+24,"frontier_days":0,"lag_months":9.5,"lag_low":6.7,"lag_high":24.7,"frontier_date":"2024-06-20","frontier_model":"Claude 3.5 Sonnet","at_frontier":false},{"model":"GPT-4.1 nano","score":129.63,"ci_low":124.26,"ci_high":131.97,"date":"2025-04-14","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":153,"training_compute_flop":null,"frontier_days":0,"lag_months":9.8,"lag_low":7.0,"lag_high":25.0,"frontier_date":"2024-06-20","frontier_model":"Claude 3.5 Sonnet","at_frontier":false,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"model":"Gemini 1.5 Flash (Sep 2024)","score":129.36,"ci_low":124.24,"ci_high":131.19,"date":"2024-09-24","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":154,"training_compute_flop":null,"frontier_days":0,"lag_months":3.2,"lag_low":0.4,"lag_high":18.4,"frontier_date":"2024-06-20","frontier_model":"Claude 3.5 Sonnet","at_frontier":false},{"model":"Qwen2.5-72B","score":129.0,"ci_low":124.2,"ci_high":130.83,"date":"2024-09-19","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":155,"training_compute_flop":7.8e+24,"frontier_days":0,"lag_months":3.0,"lag_low":0.2,"lag_high":18.2,"frontier_date":"2024-06-20","frontier_model":"Claude 3.5 Sonnet","at_frontier":false},{"model":"GPT-4o (May 2024)","score":128.97,"ci_low":125.32,"ci_high":131.55,"date":"2024-05-13","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":156,"training_compute_flop":null,"frontier_days":38,"lag_months":0.0,"lag_low":0.0,"lag_high":14.0,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":true},{"model":"GPT-4o (Nov 2024)","score":128.81,"ci_low":124.89,"ci_high":131.18,"date":"2024-11-20","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":157,"training_compute_flop":null,"frontier_days":0,"lag_months":6.3,"lag_low":2.3,"lag_high":20.3,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"GPT-4o (Aug 2024)","score":128.77,"ci_low":124.75,"ci_high":131.04,"date":"2024-08-06","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":158,"training_compute_flop":null,"frontier_days":0,"lag_months":2.8,"lag_low":2.8,"lag_high":16.8,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Llama 3.1-405B","score":128.75,"ci_low":124.83,"ci_high":130.75,"date":"2024-07-23","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":159,"training_compute_flop":3.8e+25,"frontier_days":0,"lag_months":2.3,"lag_low":2.3,"lag_high":16.3,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Qwen2.5-32B","score":128.52,"ci_low":123.46,"ci_high":130.17,"date":"2024-09-17","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":160,"training_compute_flop":3.51e+24,"frontier_days":0,"lag_months":4.2,"lag_low":0.2,"lag_high":18.2,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Mistral Large 2 (Nov 2024)","score":128.52,"ci_low":124.64,"ci_high":130.68,"date":"2024-11-18","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["France"],"rank":161,"training_compute_flop":null,"frontier_days":0,"lag_months":6.2,"lag_low":2.2,"lag_high":20.2,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Mistral Large 2 (Jul 2024)","score":127.54,"ci_low":123.38,"ci_high":129.87,"date":"2024-07-24","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["France"],"rank":162,"training_compute_flop":null,"frontier_days":0,"lag_months":2.4,"lag_low":1.1,"lag_high":16.4,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Mistral Small 3.1","score":127.48,"ci_low":123.58,"ci_high":129.41,"date":"2025-03-17","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":163,"training_compute_flop":null,"frontier_days":0,"lag_months":10.1,"lag_low":8.9,"lag_high":24.1,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"Llama 3.3 70B","score":127.32,"ci_low":123.07,"ci_high":129.81,"date":"2024-12-06","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":164,"training_compute_flop":6.8649768e+24,"frontier_days":0,"lag_months":6.8,"lag_low":5.6,"lag_high":20.8,"frontier_date":"2024-05-13","frontier_model":"GPT-4o (May 2024)","at_frontier":false},{"model":"GPT-4 Turbo (Apr 2024)","score":127.25,"ci_low":123.32,"ci_high":128.87,"date":"2024-04-09","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":165,"training_compute_flop":null,"frontier_days":34,"lag_months":0.0,"lag_low":0.0,"lag_high":12.9,"frontier_date":"2024-04-09","frontier_model":"GPT-4 Turbo (Apr 2024)","at_frontier":true},{"model":"Claude 3.5 Haiku","score":127.15,"ci_low":120.88,"ci_high":129.73,"date":"2024-10-22","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":166,"training_compute_flop":null,"frontier_days":0,"lag_months":6.4,"lag_low":4.1,"lag_high":19.3,"frontier_date":"2024-04-09","frontier_model":"GPT-4 Turbo (Apr 2024)","at_frontier":false},{"model":"Mistral Small 3","score":127.07,"ci_low":122.77,"ci_high":129.11,"date":"2025-01-30","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":167,"training_compute_flop":1.152e+24,"frontier_days":0,"lag_months":9.7,"lag_low":7.4,"lag_high":22.6,"frontier_date":"2024-04-09","frontier_model":"GPT-4 Turbo (Apr 2024)","at_frontier":false},{"model":"Claude 3 Opus","score":126.91,"ci_low":123.2,"ci_high":130.09,"date":"2024-02-29","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":168,"training_compute_flop":null,"frontier_days":40,"lag_months":0.0,"lag_low":0.0,"lag_high":11.6,"frontier_date":"2024-02-29","frontier_model":"Claude 3 Opus","at_frontier":true},{"model":"Gemini 1.5 Pro (May 2024)","score":126.9,"ci_low":122.76,"ci_high":129.18,"date":"2024-05-14","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":169,"training_compute_flop":null,"frontier_days":0,"lag_months":2.5,"lag_low":2.5,"lag_high":14.0,"frontier_date":"2024-02-29","frontier_model":"Claude 3 Opus","at_frontier":false},{"model":"GPT-4o mini","score":126.56,"ci_low":121.04,"ci_high":128.51,"date":"2024-07-18","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":170,"training_compute_flop":null,"frontier_days":0,"lag_months":4.6,"lag_low":2.2,"lag_high":16.2,"frontier_date":"2024-02-29","frontier_model":"Claude 3 Opus","at_frontier":false,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-4 Turbo (Nov 2023)","score":126.45,"ci_low":122.53,"ci_high":129.06,"date":"2024-01-25","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":171,"training_compute_flop":null,"frontier_days":35,"lag_months":0.0,"lag_low":0.0,"lag_high":10.4,"frontier_date":"2024-01-25","frontier_model":"GPT-4 Turbo (Nov 2023)","at_frontier":true},{"model":"Llama 3.1-70B","score":125.91,"ci_low":121.49,"ci_high":128.4,"date":"2024-07-23","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":172,"training_compute_flop":7.929e+24,"frontier_days":0,"lag_months":5.9,"lag_low":2.3,"lag_high":16.3,"frontier_date":"2024-01-25","frontier_model":"GPT-4 Turbo (Nov 2023)","at_frontier":false},{"model":"GPT-4 (Mar 2023)","score":125.89,"ci_low":120.86,"ci_high":130.92,"date":"2023-03-14","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":173,"training_compute_flop":2.1e+25,"frontier_days":317,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":true},{"model":"Llama 3.2 90B","score":125.5,"ci_low":120.84,"ci_high":127.41,"date":"2024-09-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":174,"training_compute_flop":null,"frontier_days":0,"lag_months":18.4,"lag_low":4.4,"lag_high":18.4,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen2-72B","score":125.27,"ci_low":120.23,"ci_high":126.96,"date":"2024-06-07","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":175,"training_compute_flop":3.02e+24,"frontier_days":0,"lag_months":14.8,"lag_low":1.9,"lag_high":14.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":124.77,"ci_low":119.95,"ci_high":127.8,"date":"2024-05-07","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":176,"training_compute_flop":1.02e+24,"frontier_days":0,"lag_months":13.8,"lag_low":13.8,"lag_high":13.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Amazon Nova Pro","score":123.77,"ci_low":109.93,"ci_high":126.64,"date":"2024-12-03","organization":"Amazon","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":177,"training_compute_flop":6.000010000000001e+24,"frontier_days":0,"lag_months":20.7,"lag_low":9.1,"lag_high":21.3,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemma 3 12B","score":123.46,"ci_low":115.68,"ci_high":129.29,"date":"2025-03-12","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":178,"training_compute_flop":8.64e+23,"frontier_days":0,"lag_months":23.9,"lag_low":8.7,"lag_high":23.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"GPT-4 (Jun 2023)","score":123.1,"ci_low":117.33,"ci_high":126.96,"date":"2023-06-13","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":179,"training_compute_flop":2.1e+25,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":3.0,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Llama 3-70B","score":122.91,"ci_low":118.22,"ci_high":125.97,"date":"2024-04-18","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":180,"training_compute_flop":7.861e+24,"frontier_days":0,"lag_months":13.2,"lag_low":2.8,"lag_high":13.2,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemini 1.5 Flash (May 2024)","score":122.57,"ci_low":117.47,"ci_high":124.66,"date":"2024-05-23","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":181,"training_compute_flop":null,"frontier_days":0,"lag_months":14.3,"lag_low":14.3,"lag_high":14.3,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemma 2 27B","score":122.05,"ci_low":116.0,"ci_high":124.19,"date":"2024-06-24","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":182,"training_compute_flop":2.106e+24,"frontier_days":0,"lag_months":15.4,"lag_low":15.4,"lag_high":15.4,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Mixtral 8x22B","score":122.0,"ci_low":115.9,"ci_high":124.51,"date":"2024-04-17","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":183,"training_compute_flop":2.34e+24,"frontier_days":0,"lag_months":13.1,"lag_low":13.1,"lag_high":13.1,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Mistral Large","score":122.0,"ci_low":115.45,"ci_high":124.63,"date":"2024-02-26","organization":"Mistral AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["France"],"rank":184,"training_compute_flop":null,"frontier_days":0,"lag_months":11.5,"lag_low":11.5,"lag_high":11.5,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"phi-3-small 7.4B","score":121.8,"ci_low":116.07,"ci_high":124.75,"date":"2024-04-23","organization":"Microsoft","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":185,"training_compute_flop":2.1312e+23,"frontier_days":0,"lag_months":13.3,"lag_low":13.3,"lag_high":13.3,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"phi-3-medium 14B","score":121.19,"ci_low":115.62,"ci_high":124.27,"date":"2024-04-23","organization":"Microsoft","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":186,"training_compute_flop":4.032e+23,"frontier_days":0,"lag_months":13.3,"lag_low":13.3,"lag_high":13.3,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Claude 3 Sonnet","score":120.68,"ci_low":113.41,"ci_high":124.15,"date":"2024-02-29","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":187,"training_compute_flop":null,"frontier_days":0,"lag_months":11.6,"lag_low":11.6,"lag_high":11.6,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Claude Instant","score":120.2,"ci_low":114.06,"ci_high":122.84,"date":"2023-08-09","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":188,"training_compute_flop":null,"frontier_days":0,"lag_months":4.9,"lag_low":4.9,"lag_high":4.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Claude 2","score":120.09,"ci_low":114.65,"ci_high":123.88,"date":"2023-07-11","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":189,"training_compute_flop":3.866e+24,"frontier_days":0,"lag_months":3.9,"lag_low":3.9,"lag_high":3.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemma 2 9B","score":119.79,"ci_low":112.51,"ci_high":122.42,"date":"2024-06-24","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":190,"training_compute_flop":4.32e+23,"frontier_days":0,"lag_months":15.4,"lag_low":15.4,"lag_high":15.4,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen2.5-Coder-32B","score":119.42,"ci_low":110.95,"ci_high":124.94,"date":"2024-09-18","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":191,"training_compute_flop":1.0725e+24,"frontier_days":0,"lag_months":18.2,"lag_low":18.2,"lag_high":18.2,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Command R+","score":119.25,"ci_low":111.34,"ci_high":125.27,"date":"2024-08-30","organization":"Cohere,Cohere Labs (formerly Cohere for AI)","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["Canada"],"rank":192,"training_compute_flop":null,"frontier_days":0,"lag_months":17.6,"lag_low":17.6,"lag_high":17.6,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Claude 2.1","score":119.22,"ci_low":113.09,"ci_high":122.07,"date":"2023-11-21","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":193,"training_compute_flop":null,"frontier_days":0,"lag_months":8.3,"lag_low":8.3,"lag_high":8.3,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Mistral NeMo","score":118.64,"ci_low":112.08,"ci_high":121.77,"date":"2024-07-18","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":194,"training_compute_flop":null,"frontier_days":0,"lag_months":16.2,"lag_low":16.2,"lag_high":16.2,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false,"model_id":"mistral/mistral-nemo","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-nemo.toml"},{"model":"GPT-3.5 Turbo (Nov 2023)","score":118.49,"ci_low":109.97,"ci_high":121.66,"date":"2023-11-06","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":195,"training_compute_flop":null,"frontier_days":0,"lag_months":7.8,"lag_low":7.8,"lag_high":7.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen2.5-7B","score":118.44,"ci_low":111.37,"ci_high":121.21,"date":"2024-09-19","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":196,"training_compute_flop":8.2188e+23,"frontier_days":0,"lag_months":18.2,"lag_low":18.2,"lag_high":18.2,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Mixtral 8x7B","score":118.4,"ci_low":112.43,"ci_high":121.19,"date":"2023-12-11","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":197,"training_compute_flop":7.74e+23,"frontier_days":0,"lag_months":8.9,"lag_low":8.9,"lag_high":8.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Claude 3 Haiku","score":118.3,"ci_low":112.27,"ci_high":121.43,"date":"2024-03-07","organization":"Anthropic","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":198,"training_compute_flop":null,"frontier_days":0,"lag_months":11.8,"lag_low":11.8,"lag_high":11.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Ministral 3B","score":118.06,"ci_low":106.88,"ci_high":121.5,"date":"2024-10-16","organization":"Mistral AI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["France"],"rank":199,"training_compute_flop":null,"frontier_days":0,"lag_months":19.1,"lag_low":19.1,"lag_high":19.7,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false,"model_id":"mistral/ministral-3b","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/ministral-3b.toml"},{"model":"Yi-34B","score":117.3,"ci_low":110.2,"ci_high":120.62,"date":"2023-11-02","organization":"01.AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":200,"training_compute_flop":6.1e+23,"frontier_days":0,"lag_months":7.7,"lag_low":7.7,"lag_high":7.7,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"phi-3-mini 3.8B","score":117.25,"ci_low":108.88,"ci_high":121.19,"date":"2024-04-23","organization":"Microsoft","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":201,"training_compute_flop":7.524e+22,"frontier_days":0,"lag_months":13.3,"lag_low":13.3,"lag_high":13.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Stable Beluga 2","score":117.0,"ci_low":110.4,"ci_high":120.04,"date":"2023-07-20","organization":"Stability AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["United Kingdom"],"rank":202,"training_compute_flop":null,"frontier_days":0,"lag_months":4.2,"lag_low":4.2,"lag_high":4.2,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemini 1.0 Pro","score":116.97,"ci_low":111.28,"ci_high":119.91,"date":"2023-12-13","organization":"Google DeepMind","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":203,"training_compute_flop":null,"frontier_days":0,"lag_months":9.0,"lag_low":9.0,"lag_high":9.0,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Llama 3.1-8B","score":116.5,"ci_low":107.07,"ci_high":121.8,"date":"2024-07-23","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":204,"training_compute_flop":1.224e+24,"frontier_days":0,"lag_months":16.3,"lag_low":16.3,"lag_high":16.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Llama 3-8B","score":116.35,"ci_low":109.0,"ci_high":119.28,"date":"2024-04-18","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":205,"training_compute_flop":7.2e+23,"frontier_days":0,"lag_months":13.2,"lag_low":13.2,"lag_high":13.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen2.5-Coder-14B","score":116.28,"ci_low":107.63,"ci_high":122.1,"date":"2024-09-18","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":206,"training_compute_flop":null,"frontier_days":0,"lag_months":18.2,"lag_low":18.2,"lag_high":18.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemma 3 4B","score":115.97,"ci_low":97.92,"ci_high":123.7,"date":"2025-03-12","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":207,"training_compute_flop":9.6e+22,"frontier_days":0,"lag_months":23.9,"lag_low":23.9,"lag_high":24.5,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"GPT-3.5 Turbo (Jan 2024)","score":115.61,"ci_low":109.31,"ci_high":119.09,"date":"2024-01-25","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":208,"training_compute_flop":null,"frontier_days":0,"lag_months":10.4,"lag_low":10.4,"lag_high":11.0,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"PaLM 2-L","score":114.88,"ci_low":107.94,"ci_high":128.04,"date":"2023-05-17","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":209,"training_compute_flop":null,"frontier_days":0,"lag_months":2.1,"lag_low":2.1,"lag_high":2.7,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Llama 2-70B","score":113.63,"ci_low":107.1,"ci_high":117.47,"date":"2023-07-18","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":210,"training_compute_flop":8.1e+23,"frontier_days":0,"lag_months":4.1,"lag_low":4.1,"lag_high":4.7,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"GPT-3.5 Turbo (Jun 2023)","score":113.19,"ci_low":105.86,"ci_high":118.59,"date":"2023-06-13","organization":"OpenAI","open_weights":false,"accessibility":"API access","accessibility_group":"Closed weights","countries":["United States of America"],"rank":211,"training_compute_flop":null,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":3.6,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen2.5-Coder (7B)","score":112.97,"ci_low":103.06,"ci_high":119.27,"date":"2024-09-18","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":212,"training_compute_flop":2.5113e+23,"frontier_days":0,"lag_months":18.2,"lag_low":18.2,"lag_high":18.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Qwen-14B","score":112.85,"ci_low":104.94,"ci_high":117.63,"date":"2023-09-24","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":213,"training_compute_flop":2.5e+23,"frontier_days":0,"lag_months":6.4,"lag_low":6.4,"lag_high":7.0,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Mistral 7B v0.1","score":112.02,"ci_low":105.4,"ci_high":116.2,"date":"2023-09-27","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":214,"training_compute_flop":null,"frontier_days":0,"lag_months":6.5,"lag_low":6.5,"lag_high":7.1,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Falcon-180B","score":111.94,"ci_low":104.28,"ci_high":118.9,"date":"2023-09-06","organization":"Technology Innovation Institute","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United Arab Emirates"],"rank":215,"training_compute_flop":3.76e+24,"frontier_days":0,"lag_months":5.8,"lag_low":5.8,"lag_high":6.4,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"internlm-20b","score":111.89,"ci_low":104.33,"ci_high":116.15,"date":"2023-09-18","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":216,"training_compute_flop":null,"frontier_days":0,"lag_months":6.2,"lag_low":6.2,"lag_high":6.8,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"Gemma 7B","score":111.79,"ci_low":104.47,"ci_high":116.73,"date":"2024-02-21","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":217,"training_compute_flop":3.07e+23,"frontier_days":0,"lag_months":11.3,"lag_low":11.3,"lag_high":11.9,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"DeepSeek LLM 67B","score":110.56,"ci_low":94.72,"ci_high":117.04,"date":"2023-11-29","organization":"DeepSeek","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":218,"training_compute_flop":8.04e+23,"frontier_days":0,"lag_months":8.5,"lag_low":8.5,"lag_high":9.1,"frontier_date":"2023-03-14","frontier_model":"GPT-4 (Mar 2023)","at_frontier":false},{"model":"LLaMA-65B","score":109.94,"ci_low":102.38,"ci_high":114.19,"date":"2023-02-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["United States of America"],"rank":219,"training_compute_flop":5.5e+23,"frontier_days":18,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":true},{"model":"Falcon 2 11B","score":109.32,"ci_low":102.33,"ci_high":115.66,"date":"2024-05-09","organization":"Technology Innovation Institute","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United Arab Emirates"],"rank":220,"training_compute_flop":3.6e+23,"frontier_days":0,"lag_months":14.5,"lag_low":13.9,"lag_high":14.5,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Mistral 7B v0.3","score":108.78,"ci_low":98.81,"ci_high":112.63,"date":"2024-05-27","organization":"Mistral AI","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["France"],"rank":221,"training_compute_flop":null,"frontier_days":0,"lag_months":15.0,"lag_low":14.5,"lag_high":15.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"DeepSeek-Coder-V2-Lite-Base","score":108.64,"ci_low":100.76,"ci_high":113.71,"date":"2024-06-13","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":222,"training_compute_flop":null,"frontier_days":0,"lag_months":15.6,"lag_low":15.0,"lag_high":15.6,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"PaLM 2-M","score":108.04,"ci_low":100.36,"ci_high":116.54,"date":"2023-05-17","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":223,"training_compute_flop":null,"frontier_days":0,"lag_months":2.7,"lag_low":2.1,"lag_high":2.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Phi-2","score":107.69,"ci_low":86.08,"ci_high":112.67,"date":"2023-12-12","organization":"Microsoft","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":224,"training_compute_flop":2.27e+22,"frontier_days":0,"lag_months":9.6,"lag_low":9.0,"lag_high":9.6,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Qwen2.5-Coder-3B","score":107.48,"ci_low":94.58,"ci_high":114.98,"date":"2024-09-18","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":225,"training_compute_flop":null,"frontier_days":0,"lag_months":18.8,"lag_low":18.2,"lag_high":18.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Nemotron-4 15B","score":107.45,"ci_low":100.02,"ci_high":113.16,"date":"2024-02-26","organization":"NVIDIA","open_weights":false,"accessibility":"Unreleased","accessibility_group":"Closed weights","countries":["United States of America"],"rank":226,"training_compute_flop":7.5005116e+23,"frontier_days":0,"lag_months":12.1,"lag_low":11.5,"lag_high":12.1,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Yi-9B","score":107.35,"ci_low":98.96,"ci_high":115.54,"date":"2024-03-01","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":227,"training_compute_flop":null,"frontier_days":0,"lag_months":12.2,"lag_low":11.6,"lag_high":12.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"LLaMA-33B","score":107.14,"ci_low":99.12,"ci_high":111.49,"date":"2023-02-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["United States of America"],"rank":228,"training_compute_flop":2.7300000000001e+23,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":true},{"model":"Qwen-7B","score":106.55,"ci_low":97.18,"ci_high":112.2,"date":"2023-09-28","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":229,"training_compute_flop":1.01e+23,"frontier_days":0,"lag_months":7.1,"lag_low":6.5,"lag_high":7.1,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Llama 2-13B","score":105.88,"ci_low":97.93,"ci_high":111.08,"date":"2023-07-18","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":230,"training_compute_flop":1.6e+23,"frontier_days":0,"lag_months":4.7,"lag_low":4.1,"lag_high":4.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"PaLM 2-S","score":105.88,"ci_low":96.89,"ci_high":114.47,"date":"2023-05-17","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":231,"training_compute_flop":null,"frontier_days":0,"lag_months":2.7,"lag_low":2.1,"lag_high":2.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Llama 2-34B","score":104.9,"ci_low":96.63,"ci_high":110.21,"date":"2023-07-18","organization":"Meta AI","open_weights":false,"accessibility":"Unreleased","accessibility_group":"Closed weights","countries":["United States of America"],"rank":232,"training_compute_flop":4.08e+23,"frontier_days":0,"lag_months":4.7,"lag_low":4.1,"lag_high":4.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"StarCoder 2 15B","score":104.7,"ci_low":94.66,"ci_high":111.42,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":233,"training_compute_flop":3.87e+23,"frontier_days":0,"lag_months":11.9,"lag_low":11.3,"lag_high":11.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Yi 6B","score":104.46,"ci_low":96.4,"ci_high":109.77,"date":"2023-11-22","organization":"01.AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":234,"training_compute_flop":1.26e+23,"frontier_days":0,"lag_months":8.9,"lag_low":8.9,"lag_high":8.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Falcon-40B","score":104.13,"ci_low":96.13,"ci_high":109.74,"date":"2023-05-25","organization":"Technology Innovation Institute","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United Arab Emirates"],"rank":235,"training_compute_flop":2.4e+23,"frontier_days":0,"lag_months":3.0,"lag_low":3.0,"lag_high":3.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Baichuan2-13B","score":102.86,"ci_low":92.28,"ci_high":109.77,"date":"2023-09-06","organization":"Baichuan","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":236,"training_compute_flop":2.03e+23,"frontier_days":0,"lag_months":6.4,"lag_low":6.4,"lag_high":6.4,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Qwen2.5-Coder (1.5B)","score":102.66,"ci_low":87.49,"ci_high":110.49,"date":"2024-09-18","organization":"Alibaba","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["China"],"rank":237,"training_compute_flop":5.082e+22,"frontier_days":0,"lag_months":18.8,"lag_low":18.2,"lag_high":18.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"internlm-7b","score":102.53,"ci_low":93.35,"ci_high":108.96,"date":"2023-07-05","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":238,"training_compute_flop":null,"frontier_days":0,"lag_months":4.3,"lag_low":4.3,"lag_high":4.3,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Llama 3.2 1B","score":102.43,"ci_low":92.05,"ci_high":113.95,"date":"2024-09-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":239,"training_compute_flop":6.642e+22,"frontier_days":0,"lag_months":19.0,"lag_low":18.4,"lag_high":19.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false,"model_id":"meta/llama-3.2-1b","lab":"meta","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/llama-3.2-1b.toml"},{"model":"INTELLECT-1","score":100.49,"ci_low":90.54,"ci_high":106.31,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":["United States of America"],"rank":240,"training_compute_flop":6.000001e+22,"frontier_days":0,"lag_months":21.2,"lag_low":21.2,"lag_high":21.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"MPT-30B","score":100.28,"ci_low":91.03,"ci_high":106.26,"date":"2023-06-22","organization":"MosaicML","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":241,"training_compute_flop":1.8900000000001e+23,"frontier_days":0,"lag_months":3.9,"lag_low":3.9,"lag_high":3.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"LLaMA-13B","score":100.2,"ci_low":90.99,"ci_high":106.42,"date":"2023-02-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["United States of America"],"rank":242,"training_compute_flop":7.8e+22,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":true},{"model":"Llama 2-7B","score":98.66,"ci_low":89.15,"ci_high":104.53,"date":"2023-07-18","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":243,"training_compute_flop":8.4e+22,"frontier_days":0,"lag_months":4.7,"lag_low":4.7,"lag_high":4.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"chatglm2-6b","score":98.53,"ci_low":80.7,"ci_high":105.99,"date":"2023-06-24","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":244,"training_compute_flop":null,"frontier_days":0,"lag_months":3.9,"lag_low":3.9,"lag_high":3.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"LLaMA-7B","score":96.25,"ci_low":86.62,"ci_high":103.51,"date":"2023-02-24","organization":"Meta AI","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["United States of America"],"rank":245,"training_compute_flop":4.00000001e+22,"frontier_days":0,"lag_months":0.0,"lag_low":0.0,"lag_high":0.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":true},{"model":"Baichuan 2-7B","score":95.88,"ci_low":85.35,"ci_high":104.45,"date":"2023-09-20","organization":"Baichuan","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":246,"training_compute_flop":1.092e+23,"frontier_days":0,"lag_months":6.8,"lag_low":6.8,"lag_high":6.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"DeepSeek Coder 33B","score":95.87,"ci_low":83.8,"ci_high":103.31,"date":"2023-11-02","organization":"DeepSeek,Peking University","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":247,"training_compute_flop":3.96e+23,"frontier_days":0,"lag_months":8.2,"lag_low":8.2,"lag_high":8.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Falcon-7B","score":94.63,"ci_low":83.39,"ci_high":102.26,"date":"2023-04-24","organization":"Technology Innovation Institute","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United Arab Emirates"],"rank":248,"training_compute_flop":6.3e+22,"frontier_days":0,"lag_months":1.9,"lag_low":1.9,"lag_high":1.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"CodeQwen1.5-7B","score":94.38,"ci_low":80.91,"ci_high":103.33,"date":"2024-04-15","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":249,"training_compute_flop":null,"frontier_days":0,"lag_months":13.7,"lag_low":13.7,"lag_high":13.7,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"vicuna-13b-v1.1","score":94.19,"ci_low":81.5,"ci_high":102.35,"date":"2023-04-12","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":250,"training_compute_flop":null,"frontier_days":0,"lag_months":1.5,"lag_low":1.5,"lag_high":1.5,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"MPT-7B","score":94.11,"ci_low":82.85,"ci_high":101.31,"date":"2023-05-05","organization":"MosaicML","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":251,"training_compute_flop":4.2000000000000004e+22,"frontier_days":0,"lag_months":2.3,"lag_low":2.3,"lag_high":2.3,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Gemma 2B","score":93.72,"ci_low":83.4,"ci_high":100.44,"date":"2024-02-21","organization":"Google DeepMind","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":252,"training_compute_flop":4.5115822e+22,"frontier_days":0,"lag_months":11.9,"lag_low":11.9,"lag_high":11.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"StarCoder 2 7B","score":93.06,"ci_low":79.66,"ci_high":101.53,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":253,"training_compute_flop":1.55e+23,"frontier_days":0,"lag_months":11.9,"lag_low":11.9,"lag_high":11.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"XGen-7B","score":92.9,"ci_low":82.02,"ci_high":100.1,"date":"2023-06-27","organization":"Salesforce","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":254,"training_compute_flop":8.02e+22,"frontier_days":0,"lag_months":4.0,"lag_low":4.0,"lag_high":4.0,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Qwen-1_8B","score":92.43,"ci_low":72.31,"ci_high":101.42,"date":"2023-11-30","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":255,"training_compute_flop":null,"frontier_days":0,"lag_months":9.2,"lag_low":9.2,"lag_high":9.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"open_llama_7b","score":91.12,"ci_low":78.31,"ci_high":98.94,"date":"2023-06-07","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":256,"training_compute_flop":null,"frontier_days":0,"lag_months":3.4,"lag_low":3.4,"lag_high":3.4,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Phi-1.5","score":90.99,"ci_low":69.51,"ci_high":100.73,"date":"2023-09-11","organization":"Microsoft","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":257,"training_compute_flop":1.17e+21,"frontier_days":0,"lag_months":6.5,"lag_low":6.5,"lag_high":6.5,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Baichuan1-7B","score":89.92,"ci_low":78.81,"ci_high":98.42,"date":"2023-06-01","organization":"Baichuan","open_weights":true,"accessibility":"Open weights (non-commercial)","accessibility_group":"Open weights","countries":["China"],"rank":258,"training_compute_flop":5.04e+22,"frontier_days":0,"lag_months":3.2,"lag_low":3.2,"lag_high":3.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"RedPajama-INCITE-7B-Base","score":89.87,"ci_low":77.76,"ci_high":97.7,"date":"2023-05-04","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":259,"training_compute_flop":null,"frontier_days":0,"lag_months":2.3,"lag_low":2.3,"lag_high":2.3,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Dolly 2.0-12b","score":89.11,"ci_low":74.2,"ci_high":97.44,"date":"2023-04-11","organization":"Databricks","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":260,"training_compute_flop":null,"frontier_days":0,"lag_months":1.5,"lag_low":1.5,"lag_high":1.5,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"DeepSeek Coder 6.7B","score":89.05,"ci_low":76.34,"ci_high":97.47,"date":"2023-11-02","organization":"DeepSeek,Peking University","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":261,"training_compute_flop":8.04e+22,"frontier_days":0,"lag_months":8.2,"lag_low":8.2,"lag_high":8.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"StarCoder 2 3B","score":88.18,"ci_low":74.45,"ci_high":97.14,"date":"2024-02-22","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["United States of America"],"rank":262,"training_compute_flop":5.94e+22,"frontier_days":0,"lag_months":11.9,"lag_low":11.9,"lag_high":11.9,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"Qwen2.5-Coder-0.5B","score":87.56,"ci_low":63.11,"ci_high":101.0,"date":"2024-09-18","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":263,"training_compute_flop":null,"frontier_days":0,"lag_months":18.8,"lag_low":18.8,"lag_high":18.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false,"model_id":"alibaba/qwen2.5-coder-0.5b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-0.5b.toml"},{"model":"Cerebras-GPT-13B","score":82.57,"ci_low":68.78,"ci_high":91.55,"date":"2023-03-20","organization":"Cerebras Systems","open_weights":true,"accessibility":"Open weights (unrestricted)","accessibility_group":"Open weights","countries":["United States of America"],"rank":264,"training_compute_flop":2.3e+22,"frontier_days":0,"lag_months":0.8,"lag_low":0.8,"lag_high":0.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"DeepSeek Coder 1.3B","score":62.54,"ci_low":47.12,"ci_high":77.27,"date":"2023-11-02","organization":"DeepSeek,Peking University","open_weights":true,"accessibility":"Open weights (restricted use)","accessibility_group":"Open weights","countries":["China"],"rank":265,"training_compute_flop":1.56e+22,"frontier_days":0,"lag_months":8.2,"lag_low":8.2,"lag_high":8.2,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false},{"model":"stablelm-tuned-alpha-7b","score":54.64,"ci_low":35.01,"ci_high":70.6,"date":"2023-04-19","organization":null,"open_weights":false,"accessibility":"Unknown","accessibility_group":"Other","countries":[],"rank":266,"training_compute_flop":null,"frontier_days":0,"lag_months":1.8,"lag_low":1.8,"lag_high":1.8,"frontier_date":"2023-02-24","frontier_model":"LLaMA-65B","at_frontier":false}],"joined_to_catalog":97},"boards":[{"id":"gpqa-diamond","name":"GPQA diamond","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2023-11-20","entrants":221,"leader":{"model":"GPT-6 Astra","score":0.9577,"effort":"max","percent":95.8,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.9577,"effort":"max","percent":95.8,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Gemini 3.8 Flash","score":0.9539,"effort":"high","percent":95.4,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Gemini 3.7 Flash","score":0.9482,"effort":"high","percent":94.8,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GPT-5.4 Pro","score":0.946,"effort":"xhigh","percent":94.6,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.9444,"effort":null,"percent":94.4,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Gemini 3.6 Flash","score":0.9413,"effort":null,"percent":94.1,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Grok 4.6","score":0.94,"effort":"high","percent":94.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.5","score":0.94,"effort":"xhigh","percent":94.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.5 Pro","score":0.9392,"effort":"xhigh","percent":93.9,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"Claude Opus 5","score":0.9388,"effort":"max","percent":93.9,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.6 Sol","score":0.935,"effort":"max","percent":93.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Grok 4.5","score":0.9343,"effort":"high","percent":93.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"GPT-5.6 Terra","score":0.9331,"effort":"max","percent":93.3,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.4","score":0.933,"effort":"xhigh","percent":93.3,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Kimi K3","score":0.9312,"effort":"max","percent":93.1,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":15,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.5 Flash","score":0.928,"effort":null,"percent":92.8,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Qwen3.8 Max","score":0.9268,"effort":"xhigh","percent":92.7,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":17,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Gemini 3 Pro Preview","score":0.9261,"effort":null,"percent":92.6,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Qwen3.8 Max (0902)","score":0.923,"effort":"xhigh","percent":92.3,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":19},{"model":"GLM-5.2","score":0.9186,"effort":null,"percent":91.9,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":20,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.9167,"effort":"max","percent":91.7,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"GPT-5.6 Luna","score":0.916,"effort":"max","percent":91.6,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"GPT-5.2","score":0.914,"effort":"xhigh","percent":91.4,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Opus 4.8","score":0.9104,"effort":null,"percent":91.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.9104,"effort":"max","percent":91.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":25,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"GLM-5.3","score":0.9091,"effort":"max","percent":90.9,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":26,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"MiniMax-M3","score":0.9091,"effort":null,"percent":90.9,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":27},{"model":"Qwen3.7 Max","score":0.9091,"effort":null,"percent":90.9,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":28,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"DeepSeek v4 Pro","score":0.9091,"effort":"high","percent":90.9,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Kimi K2.6","score":0.9078,"effort":null,"percent":90.8,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Claude Sonnet 5","score":0.9053,"effort":"xhigh","percent":90.5,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Claude Opus 4.6 (32k thinking)","score":0.9053,"effort":null,"percent":90.5,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"GLM-5.3-Flash","score":0.9015,"effort":"max","percent":90.2,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Claude Opus 4.7","score":0.9015,"effort":"xhigh","percent":90.2,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GLM-5.1","score":0.899,"effort":null,"percent":89.9,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":35,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Muse Spark","score":0.898,"effort":null,"percent":89.8,"date":"2026-04-08","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"Gemini 3 Flash Preview","score":0.8939,"effort":null,"percent":89.4,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Grok 4.20","score":0.8933,"effort":null,"percent":89.3,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"Grok 4.3 Beta","score":0.8883,"effort":null,"percent":88.8,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":39},{"model":"Claude Opus 4.6 (64k thinking)","score":0.8876,"effort":null,"percent":88.8,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Inkling Small","score":0.8851,"effort":"xhigh","percent":88.5,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.8838,"effort":null,"percent":88.4,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":42},{"model":"Claude Opus 4.6","score":0.8838,"effort":"max","percent":88.4,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Inkling","score":0.8826,"effort":"xhigh","percent":88.3,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":44},{"model":"Kimi K2.7 Code","score":0.8788,"effort":null,"percent":87.9,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":45,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Qwen3.7 Plus","score":0.8788,"effort":null,"percent":87.9,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":46,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"GLM-5","score":0.8782,"effort":null,"percent":87.8,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5.1","score":0.8763,"effort":"high","percent":87.6,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Kimi K2.5 (Fireworks)","score":0.876,"effort":null,"percent":87.6,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":49},{"model":"Qwen 3.6 Max (Preview)","score":0.8737,"effort":null,"percent":87.4,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":50},{"model":"Claude Sonnet 4.6 (32k thinking)","score":0.8737,"effort":null,"percent":87.4,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":51},{"model":"Grok 4","score":0.87,"effort":null,"percent":87.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"GPT-5.4 mini","score":0.8687,"effort":"xhigh","percent":86.9,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":53,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Qwen3.5 397B-A17B","score":0.8636,"effort":null,"percent":86.4,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":54,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"GPT-5","score":0.8617,"effort":"high","percent":86.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.8605,"effort":null,"percent":86.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"Qwen3.6 27B","score":0.8586,"effort":null,"percent":85.9,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":57,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"Claude Fable 5","score":0.8586,"effort":"max","percent":85.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":58,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 4.5 (16k thinking)","score":0.8548,"effort":null,"percent":85.5,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"Nemotron 3 Ultra","score":0.8535,"effort":null,"percent":85.4,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":60},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.8529,"effort":null,"percent":85.3,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":61},{"model":"Qwen3.5 Plus","score":0.8485,"effort":null,"percent":84.8,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":62,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Qwen3.6 35B-A3B","score":0.8485,"effort":null,"percent":84.8,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":63,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.8485,"effort":null,"percent":84.8,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"Kimi K2 Thinking Turbo","score":0.8422,"effort":null,"percent":84.2,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":65,"model_id":"moonshotai/kimi-k2-thinking-turbo","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking-turbo.toml"},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.8384,"effort":null,"percent":83.8,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":66},{"model":"Qwen3.5-35B-A3B","score":0.8346,"effort":null,"percent":83.5,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":67,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"DeepSeek-V3.2","score":0.8342,"effort":"thinking","percent":83.4,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":68,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.8333,"effort":null,"percent":83.3,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":69},{"model":"Gemini 3.5 Flash-Lite","score":0.8333,"effort":null,"percent":83.3,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":70,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Claude Sonnet 4.6","score":0.8333,"effort":"medium","percent":83.3,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":71,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"GLM-4.7","score":0.8333,"effort":null,"percent":83.3,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":72,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"GPT-5.5 Instant","score":0.8251,"effort":null,"percent":82.5,"date":"2026-05-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"Qwen3.7 Flash","score":0.8232,"effort":null,"percent":82.3,"date":"2026-07-27","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":74,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"Qwen3.5 Flash","score":0.8232,"effort":null,"percent":82.3,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":75,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"Claude Sonnet 4.5 (59k thinking)","score":0.8232,"effort":null,"percent":82.3,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":76},{"model":"o3","score":0.8182,"effort":"high","percent":81.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.8182,"effort":null,"percent":81.8,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":78,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.8173,"effort":null,"percent":81.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":79},{"model":"Claude Opus 4.5 (no thinking)","score":0.8068,"effort":null,"percent":80.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":80},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.8005,"effort":null,"percent":80.1,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":81},{"model":"Claude 3.7 Sonnet (64k thinking)","score":0.7973,"effort":null,"percent":79.7,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":82},{"model":"o4-mini","score":0.7961,"effort":"high","percent":79.6,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":83,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Sonnet 4","score":0.7919,"effort":null,"percent":79.2,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":84,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Qwen3.5 9B","score":0.7898,"effort":null,"percent":79.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":85,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.7879,"effort":null,"percent":78.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":86},{"model":"GPT-5.4 nano","score":0.7847,"effort":"high","percent":78.5,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":87,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"GPT-5.5 (no thinking)","score":0.7727,"effort":null,"percent":77.3,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":88},{"model":"Claude Opus 4.1","score":0.7727,"effort":null,"percent":77.3,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":89,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"o3-mini","score":0.7702,"effort":"high","percent":77.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":90,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"o1","score":0.7677,"effort":"high","percent":76.8,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":91,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude 3.7 Sonnet (32k thinking)","score":0.7677,"effort":null,"percent":76.8,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":92},{"model":"Claude 3.7 Sonnet (16k thinking)","score":0.7677,"effort":null,"percent":76.8,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":93},{"model":"DeepSeek-R1 (May 2025)","score":0.7633,"effort":null,"percent":76.3,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":94},{"model":"Grok-3 mini","score":0.7626,"effort":null,"percent":76.3,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":95},{"model":"Claude Opus 4","score":0.7626,"effort":null,"percent":76.3,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":96,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Gemma 4 31B IT","score":0.7576,"effort":null,"percent":75.8,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":97,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"gpt-oss-120b","score":0.7576,"effort":null,"percent":75.8,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":98,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Grok 3","score":0.7576,"effort":null,"percent":75.8,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":99},{"model":"GPT-5 mini","score":0.75,"effort":"high","percent":75.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":100,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"}]},{"id":"otis-mock-aime-2024-2025","name":"OTIS Mock AIME 2024-2025","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.001,"released":"2024-12-19","entrants":201,"leader":{"model":"Qwen3.8 Max (0902)","score":1.0,"effort":"xhigh","percent":100.0,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":1},"rows":[{"model":"Qwen3.8 Max (0902)","score":1.0,"effort":"xhigh","percent":100.0,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":1},{"model":"Claude Fable 5.1","score":1.0,"effort":"max","percent":100.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-6 Astra","score":1.0,"effort":"max","percent":100.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"GPT-5.6 Sol","score":1.0,"effort":"max","percent":100.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.5","score":1.0,"effort":"xhigh","percent":100.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Fable 5","score":1.0,"effort":"high","percent":100.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.5 Pro","score":1.0,"effort":"xhigh","percent":100.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"GPT-5.6 Terra","score":0.9972,"effort":"max","percent":99.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Qwen3.8 Max","score":0.9944,"effort":"xhigh","percent":99.4,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":9,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Grok 4.6","score":0.9917,"effort":"xhigh","percent":99.2,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Gemini 3.8 Flash","score":0.9889,"effort":"high","percent":98.9,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Claude Opus 5","score":0.9889,"effort":"max","percent":98.9,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.9861,"effort":"max","percent":98.6,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":13,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"GPT-5.6 Luna","score":0.9833,"effort":"max","percent":98.3,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Opus 4.8","score":0.9833,"effort":null,"percent":98.3,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Claude Opus 4.7","score":0.978,"effort":"xhigh","percent":97.8,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.4","score":0.9778,"effort":"high","percent":97.8,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Grok 4.5","score":0.9778,"effort":"high","percent":97.8,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Gemini 3.7 Flash","score":0.9722,"effort":"high","percent":97.2,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Kimi K3","score":0.9722,"effort":"max","percent":97.2,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":20,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"DeepSeek v4 Pro","score":0.9667,"effort":"max","percent":96.7,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5.2","score":0.9611,"effort":"high","percent":96.1,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Kimi K2.6","score":0.9611,"effort":null,"percent":96.1,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.956,"effort":null,"percent":95.6,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Qwen3.7 Max","score":0.9556,"effort":null,"percent":95.6,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":25,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Kimi K2.7 Code","score":0.9556,"effort":null,"percent":95.6,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Gemini 3.5 Flash","score":0.9556,"effort":null,"percent":95.6,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Gemini 3 Flash Preview","score":0.9556,"effort":null,"percent":95.6,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Claude Sonnet 5","score":0.9472,"effort":"xhigh","percent":94.7,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.9444,"effort":"max","percent":94.4,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Claude Opus 4.6 (64k thinking)","score":0.9444,"effort":null,"percent":94.4,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"Gemini 3.6 Flash","score":0.9417,"effort":null,"percent":94.2,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GLM-5.3-Flash","score":0.9389,"effort":"max","percent":93.9,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"GLM-5.1","score":0.9333,"effort":null,"percent":93.3,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":34,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.9333,"effort":null,"percent":93.3,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":35},{"model":"Qwen3.7 Plus","score":0.9333,"effort":null,"percent":93.3,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":36,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Grok 4.3 Beta","score":0.9333,"effort":null,"percent":93.3,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":37},{"model":"Claude Opus 4.6 (32k thinking)","score":0.9306,"effort":null,"percent":93.1,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"Grok 4.20","score":0.9222,"effort":null,"percent":92.2,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"Kimi K2.5 (Fireworks)","score":0.922,"effort":null,"percent":92.2,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":40},{"model":"GPT-5","score":0.9139,"effort":"high","percent":91.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Gemini 3 Pro Preview","score":0.9139,"effort":null,"percent":91.4,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":42,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GLM-5.3","score":0.9111,"effort":"max","percent":91.1,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":43,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Qwen 3.6 Max (Preview)","score":0.9111,"effort":null,"percent":91.1,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":44},{"model":"Qwen3.6 27B","score":0.9111,"effort":null,"percent":91.1,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":45,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"Claude Opus 4.6","score":0.9111,"effort":"max","percent":91.1,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Inkling Small","score":0.9,"effort":"xhigh","percent":90.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":47},{"model":"Muse Spark","score":0.889,"effort":null,"percent":88.9,"date":"2026-04-08","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"GPT-5.4 mini","score":0.8889,"effort":"xhigh","percent":88.9,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Qwen3.5 397B-A17B","score":0.8889,"effort":null,"percent":88.9,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":50,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"Inkling","score":0.8889,"effort":"xhigh","percent":88.9,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":51},{"model":"gpt-oss-120b","score":0.8889,"effort":null,"percent":88.9,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":52,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"GPT-5.1","score":0.8861,"effort":"high","percent":88.6,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":53,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"DeepSeek-V3.2","score":0.8782,"effort":"thinking","percent":87.8,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":54,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"model":"GPT-5.4 nano","score":0.8778,"effort":"high","percent":87.8,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Nemotron 3 Ultra","score":0.8667,"effort":null,"percent":86.7,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":56},{"model":"Qwen3.5 Plus","score":0.8667,"effort":null,"percent":86.7,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":57,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Qwen3.6 35B-A3B","score":0.8667,"effort":null,"percent":86.7,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":58,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Qwen3.7 Flash","score":0.8667,"effort":null,"percent":86.7,"date":"2026-07-27","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":59,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"GPT-5 mini","score":0.8667,"effort":"high","percent":86.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":60,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.8667,"effort":null,"percent":86.7,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":61},{"model":"GLM-5.2","score":0.8639,"effort":null,"percent":86.4,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":62,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.8611,"effort":null,"percent":86.1,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":63},{"model":"Claude Sonnet 4.6 (32k thinking)","score":0.8579,"effort":null,"percent":85.8,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.8472,"effort":null,"percent":84.7,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":65},{"model":"o3","score":0.8444,"effort":"medium","percent":84.4,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":66,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.8444,"effort":null,"percent":84.4,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":67},{"model":"Qwen3.5 Flash","score":0.8444,"effort":null,"percent":84.4,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":68,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"Grok 4","score":0.84,"effort":null,"percent":84.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":69},{"model":"GLM-4.7","score":0.8333,"effort":null,"percent":83.3,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":70,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"Kimi K2 Thinking Turbo","score":0.8306,"effort":null,"percent":83.1,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":71,"model_id":"moonshotai/kimi-k2-thinking-turbo","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking-turbo.toml"},{"model":"Gemma 4 26B A4B","score":0.8222,"effort":null,"percent":82.2,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":72},{"model":"Claude Sonnet 4.6","score":0.8222,"effort":"medium","percent":82.2,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":73,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"o4-mini","score":0.8167,"effort":"high","percent":81.7,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":74,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Opus 4.5 (16k thinking)","score":0.8167,"effort":null,"percent":81.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":75},{"model":"GPT-5 nano","score":0.8111,"effort":"high","percent":81.1,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":76,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.8,"effort":null,"percent":80.0,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":77,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"GLM-5","score":0.8,"effort":null,"percent":80.0,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":78,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"Claude Sonnet 4.5 (59k thinking)","score":0.7778,"effort":null,"percent":77.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":79},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.7778,"effort":null,"percent":77.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":80},{"model":"Grok-3 mini","score":0.7778,"effort":null,"percent":77.8,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":81},{"model":"o3-mini","score":0.7694,"effort":"high","percent":76.9,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":82,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Gemma 4 31B IT","score":0.7333,"effort":null,"percent":73.3,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":83,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"o1","score":0.7333,"effort":"medium","percent":73.3,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":84,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Qwen3-Max-Instruct","score":0.7333,"effort":null,"percent":73.3,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":85},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":0.7306,"effort":null,"percent":73.1,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":86},{"model":"MiniMax-M3","score":0.7111,"effort":null,"percent":71.1,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":87},{"model":"Gemini 3.5 Flash-Lite","score":0.7111,"effort":null,"percent":71.1,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":88,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.7111,"effort":null,"percent":71.1,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":89},{"model":"Claude Sonnet 4","score":0.7111,"effort":null,"percent":71.1,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":90,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":0.7083,"effort":null,"percent":70.8,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":91},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","score":0.7028,"effort":null,"percent":70.3,"date":"2025-07-30","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":92},{"model":"Qwen3.5-35B-A3B","score":0.7,"effort":null,"percent":70.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":93,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"Claude Opus 4.1","score":0.6889,"effort":null,"percent":68.9,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":94,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"GPT-5.5 Instant","score":0.6806,"effort":null,"percent":68.1,"date":"2026-05-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":95,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"seed-oss-36b-instruct","score":0.675,"effort":null,"percent":67.5,"date":"2025-08-20","organization":null,"countries":[],"accessibility":null,"rank":96},{"model":"Qwen3-32B","score":0.6694,"effort":null,"percent":66.9,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":97},{"model":"Claude Haiku 4.5","score":0.6667,"effort":null,"percent":66.7,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":98,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Qwen3-14B","score":0.6639,"effort":null,"percent":66.4,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":99},{"model":"DeepSeek-R1 (May 2025)","score":0.6639,"effort":null,"percent":66.4,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":100}]},{"id":"dtbench","name":"DTBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.4,"released":"2026-08-12","entrants":161,"leader":{"model":"Claude Fable 5","score":0.984,"effort":"max","percent":98.4,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.984,"effort":"max","percent":98.4,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":0.976,"effort":"max","percent":97.6,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Grok 4.6","score":0.9733,"effort":"xhigh","percent":97.3,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Gemini 3.7 Flash","score":0.968,"effort":"high","percent":96.8,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Grok 4.5","score":0.9653,"effort":"high","percent":96.5,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":0.96,"effort":null,"percent":96.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"GPT-5.5","score":0.96,"effort":"xhigh","percent":96.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.5 Pro","score":0.96,"effort":"xhigh","percent":96.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"GPT-5.6 Sol","score":0.9547,"effort":"max","percent":95.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.9547,"effort":null,"percent":95.5,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Gemini 3.6 Flash","score":0.9547,"effort":null,"percent":95.5,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Claude Opus 4.8","score":0.9493,"effort":null,"percent":94.9,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Muse Spark 1.2","score":0.9467,"effort":"xhigh","percent":94.7,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"Gemini 3.5 Flash","score":0.9467,"effort":null,"percent":94.7,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Claude Opus 4.7","score":0.9465,"effort":"max","percent":94.7,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.4","score":0.944,"effort":"xhigh","percent":94.4,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Muse Spark 1.1","score":0.944,"effort":"high","percent":94.4,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"GLM-5.2","score":0.936,"effort":null,"percent":93.6,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"GPT-5.6 Terra","score":0.9253,"effort":"max","percent":92.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Sonnet 5","score":0.9248,"effort":"max","percent":92.5,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Qwen3.7 Max","score":0.9227,"effort":null,"percent":92.3,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":21,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Qwen3.8 Max","score":0.92,"effort":"xhigh","percent":92.0,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":22,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Claude Opus 4.6","score":0.912,"effort":"max","percent":91.2,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Kimi K3","score":0.912,"effort":"max","percent":91.2,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":24,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.2","score":0.9093,"effort":"xhigh","percent":90.9,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.9093,"effort":"max","percent":90.9,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Kimi K2.6","score":0.9093,"effort":null,"percent":90.9,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"DeepSeek v4 Pro","score":0.9067,"effort":"max","percent":90.7,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":28,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5","score":0.9067,"effort":"high","percent":90.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Grok 4.3 Beta","score":0.9067,"effort":null,"percent":90.7,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":30},{"model":"GPT-5.1","score":0.9013,"effort":"high","percent":90.1,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Grok 4.20","score":0.9013,"effort":null,"percent":90.1,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"Nemotron 3 Ultra","score":0.9013,"effort":null,"percent":90.1,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":33},{"model":"Claude Sonnet 4.6","score":0.8987,"effort":"max","percent":89.9,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Opus 4.5","score":0.8987,"effort":"high","percent":89.9,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"GPT-5.6 Luna","score":0.8907,"effort":"max","percent":89.1,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Gemini 3 Flash Preview","score":0.8907,"effort":null,"percent":89.1,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Grok 4.1 Fast","score":0.8773,"effort":null,"percent":87.7,"date":"2025-11-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"xai/grok-4.1-fast","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.1-fast.toml"},{"model":"DeepSeek-V3.2-Exp","score":0.8773,"effort":null,"percent":87.7,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":39},{"model":"Inkling","score":0.8747,"effort":"xhigh","percent":87.5,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":40},{"model":"Qwen3.5 397B-A17B","score":0.8747,"effort":null,"percent":87.5,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"Qwen 3.6 Max (Preview)","score":0.872,"effort":null,"percent":87.2,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":42},{"model":"o3-pro","score":0.8693,"effort":null,"percent":86.9,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"DeepSeek v4 Flash","score":0.864,"effort":"max","percent":86.4,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.856,"effort":null,"percent":85.6,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":45},{"model":"o3","score":0.848,"effort":"high","percent":84.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"MiMo-V2.5-Pro","score":0.8453,"effort":null,"percent":84.5,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47},{"model":"Qwen3.5-122B-A10B","score":0.8427,"effort":null,"percent":84.3,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":48,"model_id":"alibaba/qwen3.5-122b-a10b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-122b-a10b.toml"},{"model":"Qwen3.7 Plus","score":0.84,"effort":null,"percent":84.0,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":49,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.8347,"effort":null,"percent":83.5,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Claude Sonnet 4.5","score":0.832,"effort":"unknown thinking","percent":83.2,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":51,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Qwen3.5 Flash","score":0.8293,"effort":null,"percent":82.9,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":52,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"Grok 4 Fast","score":0.8267,"effort":null,"percent":82.7,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"Gemma 4 31B IT","score":0.8267,"effort":null,"percent":82.7,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":54,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"DeepSeek-V3.1","score":0.8267,"effort":null,"percent":82.7,"date":"2025-08-21","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":55,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.824,"effort":null,"percent":82.4,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"Qwen3.5 27B","score":0.824,"effort":null,"percent":82.4,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":57,"model_id":"alibaba/qwen3.5-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-27b.toml"},{"model":"Qwen3-Max-Instruct","score":0.8213,"effort":null,"percent":82.1,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":58},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.8187,"effort":null,"percent":81.9,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":59},{"model":"Claude Opus 4","score":0.816,"effort":null,"percent":81.6,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":60,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"DeepSeek-V3.1-Terminus","score":0.8133,"effort":null,"percent":81.3,"date":"2025-09-22","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":61},{"model":"Qwen Plus (Apr 2025)","score":0.8107,"effort":null,"percent":81.1,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":62},{"model":"GPT-5 mini","score":0.8053,"effort":"high","percent":80.5,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":63,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Qwen3.5 Plus","score":0.8053,"effort":null,"percent":80.5,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":64,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"GPT-5.4 nano","score":0.8027,"effort":"xhigh","percent":80.3,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":65,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.8027,"effort":null,"percent":80.3,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":66},{"model":"GPT-5.4 mini","score":0.8,"effort":"xhigh","percent":80.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":67,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Claude Opus 4.1","score":0.8,"effort":"unknown thinking","percent":80.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Qwen3.5-35B-A3B","score":0.8,"effort":null,"percent":80.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":69,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"MiniMax-M3","score":0.7893,"effort":null,"percent":78.9,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":70},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":0.784,"effort":null,"percent":78.4,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":71},{"model":"Qwen3.6 27B","score":0.7813,"effort":null,"percent":78.1,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":72,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"o4-mini","score":0.776,"effort":"high","percent":77.6,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.7707,"effort":null,"percent":77.1,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":74},{"model":"Claude Sonnet 4","score":0.7707,"effort":"unknown thinking","percent":77.1,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":75,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.768,"effort":null,"percent":76.8,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":76,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Gemini 2.5 Flash (Jun 2025)","score":0.7653,"effort":null,"percent":76.5,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"gpt-oss-120b","score":0.7627,"effort":null,"percent":76.3,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":78,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Qwen3-235B-A22B","score":0.7573,"effort":null,"percent":75.7,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":79},{"model":"Mistral Medium 3.5","score":0.7547,"effort":null,"percent":75.5,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":80,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"Gemma 4 26B A4B","score":0.7493,"effort":null,"percent":74.9,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":81},{"model":"o1","score":0.7467,"effort":"high","percent":74.7,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":82,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Qwen3.6 35B-A3B","score":0.7387,"effort":null,"percent":73.9,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":83,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Claude Haiku 4.5","score":0.736,"effort":"unknown thinking","percent":73.6,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":84,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Qwen3.5 9B","score":0.712,"effort":null,"percent":71.2,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":85,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"Mistral Small 4","score":0.7093,"effort":null,"percent":70.9,"date":"2026-03-16","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":86,"model_id":"mistral/mistral-small-2603","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-small-2603.toml"},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","score":0.6933,"effort":null,"percent":69.3,"date":"2025-07-30","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":87},{"model":"o3-mini","score":0.688,"effort":"high","percent":68.8,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":88,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"GPT-4.1 mini","score":0.688,"effort":null,"percent":68.8,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":89,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"GPT-4.1","score":0.6827,"effort":null,"percent":68.3,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":90,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"gpt-oss-20b","score":0.68,"effort":null,"percent":68.0,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":91,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.6784,"effort":null,"percent":67.8,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":92},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.6777,"effort":null,"percent":67.8,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":93},{"model":"Qwen3-32B","score":0.6747,"effort":null,"percent":67.5,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":94},{"model":"Mistral Medium 3.1","score":0.6747,"effort":null,"percent":67.5,"date":"2025-08-12","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":95},{"model":"Qwen3-30B-A3B-Instruct (Jul 2025)","score":0.672,"effort":null,"percent":67.2,"date":"2025-07-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":96},{"model":"Grok-2 (Dec 2024)","score":0.6521,"effort":null,"percent":65.2,"date":"2024-12-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":97},{"model":"Mistral Large 3","score":0.6507,"effort":null,"percent":65.1,"date":"2025-12-02","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":98,"model_id":"mistral/mistral-large-2512","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-large-2512.toml"},{"model":"DeepSeek-V3 (Mar 2025)","score":0.648,"effort":null,"percent":64.8,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":99},{"model":"GPT-4o (May 2024)","score":0.6453,"effort":null,"percent":64.5,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":100}]},{"id":"chess-puzzles","name":"Chess Puzzles","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0496,"released":"2025-12-10","entrants":145,"leader":{"model":"GPT-6 Astra","score":0.72,"effort":"max","percent":72.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.72,"effort":"max","percent":72.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":0.64,"effort":null,"percent":64.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"GPT-5.5 Pro","score":0.64,"effort":"xhigh","percent":64.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"Gemini 3.8 Flash","score":0.61,"effort":"high","percent":61.0,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"GPT-5.4 Pro","score":0.586,"effort":"xhigh","percent":58.6,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"GPT-5.6 Sol","score":0.55,"effort":"max","percent":55.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.55,"effort":null,"percent":55.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.6 Terra","score":0.54,"effort":"max","percent":54.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.5","score":0.54,"effort":"xhigh","percent":54.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.5 Flash","score":0.5,"effort":null,"percent":50.0,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"GPT-5.2","score":0.49,"effort":"xhigh","percent":49.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Fable 5.1","score":0.47,"effort":"max","percent":47.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.47,"effort":"max","percent":47.0,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":13,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Gemini 3.7 Flash","score":0.47,"effort":"high","percent":47.0,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GPT-5.4","score":0.44,"effort":"xhigh","percent":44.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.6 Flash","score":0.43,"effort":null,"percent":43.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Claude Opus 5","score":0.42,"effort":"max","percent":42.0,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Fable 5","score":0.41,"effort":"high","percent":41.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Qwen3.8 Max (0902)","score":0.4,"effort":"xhigh","percent":40.0,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":19},{"model":"Grok 4.6","score":0.4,"effort":"high","percent":40.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.6 Luna","score":0.4,"effort":"max","percent":40.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Gemini 3 Flash Preview","score":0.4,"effort":null,"percent":40.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Kimi K3","score":0.39,"effort":"max","percent":39.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":23,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"o3","score":0.38,"effort":"medium","percent":38.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"GPT-5","score":0.37,"effort":"high","percent":37.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Grok 4.5","score":0.36,"effort":"high","percent":36.0,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Claude Sonnet 5","score":0.35,"effort":"xhigh","percent":35.0,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Claude Opus 4.8","score":0.34,"effort":null,"percent":34.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.33,"effort":"max","percent":33.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"GPT-5.1","score":0.32,"effort":"high","percent":32.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Gemini 3 Pro Preview","score":0.31,"effort":null,"percent":31.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5 mini","score":0.3,"effort":"high","percent":30.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GPT-5.4 nano","score":0.3,"effort":"high","percent":30.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Claude Opus 4.7","score":0.3,"effort":"xhigh","percent":30.0,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Qwen3.8 Max","score":0.29,"effort":"xhigh","percent":29.0,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":35,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Grok 4","score":0.28,"effort":null,"percent":28.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"GPT-5 nano","score":0.27,"effort":"high","percent":27.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"o4-mini","score":0.26,"effort":"high","percent":26.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Qwen3.6 35B-A3B","score":0.26,"effort":null,"percent":26.0,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":39,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Kimi K2.6","score":0.26,"effort":null,"percent":26.0,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":40,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.25,"effort":null,"percent":25.0,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":41,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Grok 4.3 Beta","score":0.25,"effort":null,"percent":25.0,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":42},{"model":"GPT-5.4 mini","score":0.24,"effort":"xhigh","percent":24.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Qwen3.7 Plus","score":0.24,"effort":null,"percent":24.0,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":44,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Grok 4.20","score":0.24,"effort":null,"percent":24.0,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":45},{"model":"Qwen3.7 Flash","score":0.23,"effort":null,"percent":23.0,"date":"2026-07-27","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":46,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"Qwen3.6 27B","score":0.22,"effort":null,"percent":22.0,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"Qwen3.5 Plus","score":0.22,"effort":null,"percent":22.0,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":48,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.22,"effort":null,"percent":22.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"GLM-5.3","score":0.21,"effort":"max","percent":21.0,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":50,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"GLM-5.2","score":0.21,"effort":null,"percent":21.0,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":51,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Kimi K2.7 Code","score":0.21,"effort":null,"percent":21.0,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":52,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Qwen3.5 Flash","score":0.21,"effort":null,"percent":21.0,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":53,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"Inkling","score":0.21,"effort":"xhigh","percent":21.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":54},{"model":"Qwen 3.6 Max (Preview)","score":0.2,"effort":null,"percent":20.0,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":55},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.2,"effort":null,"percent":20.0,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":56},{"model":"DeepSeek v4 Pro","score":0.2,"effort":"max","percent":20.0,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":57,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"gpt-oss-120b","score":0.2,"effort":null,"percent":20.0,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":58,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Kimi K2 Thinking Turbo","score":0.2,"effort":null,"percent":20.0,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":59,"model_id":"moonshotai/kimi-k2-thinking-turbo","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking-turbo.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.2,"effort":null,"percent":20.0,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":60},{"model":"GLM-5.1","score":0.19,"effort":null,"percent":19.0,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":61,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Qwen3.7 Max","score":0.19,"effort":null,"percent":19.0,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":62,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Inkling Small","score":0.18,"effort":"xhigh","percent":18.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":63},{"model":"GPT-5.1 (no thinking)","score":0.17,"effort":null,"percent":17.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.17,"effort":null,"percent":17.0,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":65},{"model":"o3-mini","score":0.17,"effort":"high","percent":17.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":66,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Claude Opus 4.6 (32k thinking)","score":0.17,"effort":null,"percent":17.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":67},{"model":"o1","score":0.15,"effort":"high","percent":15.0,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"GLM-5.3-Flash","score":0.14,"effort":"max","percent":14.0,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":69,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"MiniMax-M3","score":0.14,"effort":null,"percent":14.0,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":70},{"model":"Claude Opus 4.6","score":0.14,"effort":"max","percent":14.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":71,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"DeepSeek-V3.2","score":0.14,"effort":"thinking","percent":14.0,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":72,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"model":"seed-oss-36b-instruct","score":0.13,"effort":null,"percent":13.0,"date":"2025-08-20","organization":null,"countries":[],"accessibility":null,"rank":73},{"model":"Qwen3.5 397B-A17B","score":0.13,"effort":null,"percent":13.0,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":74,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"GPT-4o (Aug 2024)","score":0.13,"effort":null,"percent":13.0,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":75},{"model":"Claude Opus 4.6 (120k thinking)","score":0.13,"effort":null,"percent":13.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":76},{"model":"Claude Sonnet 4.6 (32k thinking)","score":0.13,"effort":null,"percent":13.0,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"Qwen3.5 9B","score":0.12,"effort":null,"percent":12.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":78,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"Nemotron 3 Ultra","score":0.12,"effort":null,"percent":12.0,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":79},{"model":"GPT-5.5 Instant","score":0.12,"effort":null,"percent":12.0,"date":"2026-05-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":80,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"Kimi K2.5 (Fireworks)","score":0.12,"effort":null,"percent":12.0,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":81},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.12,"effort":null,"percent":12.0,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":82},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.12,"effort":null,"percent":12.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":83},{"model":"Claude Opus 4.5 (32k thinking)","score":0.12,"effort":null,"percent":12.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":84},{"model":"Qwen3.5-35B-A3B","score":0.1,"effort":null,"percent":10.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":85,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"GPT-5.5 (no thinking)","score":0.1,"effort":null,"percent":10.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":86},{"model":"GLM-5","score":0.1,"effort":null,"percent":10.0,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":87,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"Claude Opus 4.6 (64k thinking)","score":0.1,"effort":null,"percent":10.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":88},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","score":0.08,"effort":null,"percent":8.0,"date":"2025-07-30","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":89},{"model":"Claude Sonnet 4.6","score":0.08,"effort":"medium","percent":8.0,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":90,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Haiku 4.5","score":0.08,"effort":null,"percent":8.0,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":91,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Claude Opus 4.1","score":0.07,"effort":null,"percent":7.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":92,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"GPT-4.1 mini","score":0.07,"effort":null,"percent":7.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":93,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"GPT-4.1","score":0.06,"effort":null,"percent":6.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":94,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Gemma 4 26B A4B","score":0.06,"effort":null,"percent":6.0,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":95},{"model":"GPT-4 Turbo (Apr 2024)","score":0.06,"effort":null,"percent":6.0,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":96},{"model":"GLM-4.7","score":0.06,"effort":null,"percent":6.0,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":97,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"Qwen3-32B","score":0.05,"effort":null,"percent":5.0,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":98},{"model":"QwQ-32B","score":0.05,"effort":null,"percent":5.0,"date":"2025-03-05","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":99,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"model":"Qwen3-8B","score":0.05,"effort":null,"percent":5.0,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":100}]},{"id":"weirdml","name":"WeirdML","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-01-16","entrants":139,"leader":{"model":"Claude Fable 5.1","score":0.929,"effort":"max","percent":92.9,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},"rows":[{"model":"Claude Fable 5.1","score":0.929,"effort":"max","percent":92.9,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-6 Astra","score":0.9287,"effort":"high","percent":92.9,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5","score":0.9194,"effort":"max","percent":91.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":0.9178,"effort":"max","percent":91.8,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":0.8943,"effort":null,"percent":89.4,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"GPT-5.6 Sol","score":0.8876,"effort":"high","percent":88.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.5","score":0.8491,"effort":"xhigh","percent":84.9,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Opus 4.8","score":0.8289,"effort":null,"percent":82.9,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Kimi K3","score":0.8257,"effort":"max","percent":82.6,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":9,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.3 Codex","score":0.793,"effort":null,"percent":79.3,"date":"2026-02-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"GPT-5.6 Terra","score":0.7827,"effort":"high","percent":78.3,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Opus 4.6","score":0.7795,"effort":"high","percent":78.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.4","score":0.777,"effort":"xhigh","percent":77.7,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Claude Opus 4.7","score":0.7644,"effort":"high","percent":76.4,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Claude Opus 4.7 (no thinking)","score":0.764,"effort":null,"percent":76.4,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"GLM-5.3","score":0.754,"effort":"max","percent":75.4,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":16,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"GPT-5.2","score":0.7219,"effort":"xhigh","percent":72.2,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.7207,"effort":null,"percent":72.1,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GLM-5.2","score":0.7012,"effort":null,"percent":70.1,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Gemini 3 Pro Preview","score":0.6993,"effort":null,"percent":69.9,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Claude Sonnet 5","score":0.6878,"effort":"high","percent":68.8,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Grok 4.6","score":0.6729,"effort":"high","percent":67.3,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.5 (no thinking)","score":0.6715,"effort":null,"percent":67.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"DeepSeek V4 Pro 0813","score":0.6622,"effort":"max","percent":66.2,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":24,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Claude Sonnet 4.6","score":0.6607,"effort":"medium","percent":66.1,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.6587,"effort":null,"percent":65.9,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":26},{"model":"Claude Opus 4.5 (16k thinking)","score":0.6374,"effort":null,"percent":63.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27},{"model":"DeepSeek V4 Flash 0731","score":0.6296,"effort":"max","percent":63.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":28,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Gemini 3.5 Flash","score":0.6264,"effort":null,"percent":62.6,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Gemini 3 Flash Preview","score":0.616,"effort":null,"percent":61.6,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"GPT-5.6 Luna","score":0.6086,"effort":"high","percent":60.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"GPT-5.1","score":0.6077,"effort":"high","percent":60.8,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"GPT-5","score":0.607,"effort":"high","percent":60.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5 Pro","score":0.6039,"effort":null,"percent":60.4,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"GPT-5.4 mini","score":0.603,"effort":"high","percent":60.3,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Muse Spark 1.2","score":0.603,"effort":"xhigh","percent":60.3,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"o3-pro","score":0.5821,"effort":null,"percent":58.2,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"GPT-5.4 Pro (no thinking)","score":0.5744,"effort":null,"percent":57.4,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"GLM-5.1","score":0.571,"effort":null,"percent":57.1,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":39,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Gemini 3.6 Flash","score":0.561,"effort":null,"percent":56.1,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":40,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Kimi K2.6","score":0.5586,"effort":null,"percent":55.9,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"GPT-5-codex","score":0.5453,"effort":null,"percent":54.5,"date":"2025-09-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":42,"model_id":"openai/gpt-5-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-codex.toml"},{"model":"Kimi K2.7 Code","score":0.5412,"effort":null,"percent":54.1,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.5403,"effort":null,"percent":54.0,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"GPT-5 mini","score":0.5267,"effort":"high","percent":52.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"o4-mini","score":0.5256,"effort":"high","percent":52.6,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"o3","score":0.5242,"effort":"high","percent":52.4,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":47,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Grok 4.20","score":0.5226,"effort":null,"percent":52.3,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"Gemma 4 31B IT","score":0.5226,"effort":null,"percent":52.3,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":49,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.5219,"effort":null,"percent":52.2,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Grok 4.3 Beta","score":0.4989,"effort":null,"percent":49.9,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":51},{"model":"GPT-5.4 nano","score":0.4923,"effort":"high","percent":49.2,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"DeepSeek v4 Pro","score":0.489,"effort":"max","percent":48.9,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":53,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"gpt-oss-120b","score":0.4817,"effort":null,"percent":48.2,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":54,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"GLM-5","score":0.4817,"effort":null,"percent":48.2,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":55,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.4771,"effort":null,"percent":47.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"o1-preview","score":0.4756,"effort":null,"percent":47.6,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":57},{"model":"DeepSeek-V3.2-Speciale","score":0.4673,"effort":null,"percent":46.7,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":58},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.467,"effort":null,"percent":46.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"Grok 4.5","score":0.4643,"effort":"unknown thinking","percent":46.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":60,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Claude Sonnet 4","score":0.4611,"effort":null,"percent":46.1,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":61,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"o1","score":0.4608,"effort":"high","percent":46.1,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":62,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude Opus 4.1","score":0.4586,"effort":null,"percent":45.9,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":63,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Grok 4","score":0.4573,"effort":null,"percent":45.7,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"DeepSeek v4 Flash","score":0.4563,"effort":"max","percent":45.6,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":65,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"model":"Kimi K2.5","score":0.456,"effort":null,"percent":45.6,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":66},{"model":"Claude Haiku 4.5","score":0.454,"effort":null,"percent":45.4,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":67,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Claude Opus 4","score":0.4372,"effort":null,"percent":43.7,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Mistral Medium 3.5","score":0.4371,"effort":null,"percent":43.7,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":69,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"o3-mini","score":0.437,"effort":"high","percent":43.7,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":70,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Nemotron 3 Ultra","score":0.4345,"effort":null,"percent":43.5,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":71},{"model":"Mercury 2","score":0.432,"effort":null,"percent":43.2,"date":null,"organization":"Inception Labs","countries":["United States of America"],"accessibility":"API access","rank":72},{"model":"Grok 4 Fast","score":0.4286,"effort":null,"percent":42.9,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":73},{"model":"Kimi K2 Thinking","score":0.4279,"effort":null,"percent":42.8,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":74,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"Grok 3 mini","score":0.4258,"effort":null,"percent":42.6,"date":"2025-06-24","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":75},{"model":"Grok-3 mini","score":0.4258,"effort":null,"percent":42.6,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":76},{"model":"Gemini 2.5 Flash (Sep 2025)","score":0.4191,"effort":null,"percent":41.9,"date":"2025-09-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"DeepSeek-R1 (May 2025)","score":0.4163,"effort":null,"percent":41.6,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":78},{"model":"Qwen3-Coder-480B-A35B","score":0.4117,"effort":null,"percent":41.2,"date":"2025-07-31","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":79},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.4104,"effort":null,"percent":41.0,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":80},{"model":"Gemini 2.5 Flash (Apr 2025)","score":0.4095,"effort":null,"percent":41.0,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":81},{"model":"Gemini 2.5 Flash (May 2025)","score":0.4095,"effort":null,"percent":40.9,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":82},{"model":"gpt-oss-20b","score":0.4093,"effort":null,"percent":40.9,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":83,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"},{"model":"GLM-4.5 Thinking","score":0.4058,"effort":null,"percent":40.6,"date":"2025-08-05","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":84},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.3997,"effort":null,"percent":40.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":85},{"model":"Qwen3.5 27B","score":0.3953,"effort":null,"percent":39.5,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":86,"model_id":"alibaba/qwen3.5-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-27b.toml"},{"model":"DeepSeek-V3.2-Exp","score":0.3948,"effort":null,"percent":39.5,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":87},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.3937,"effort":null,"percent":39.4,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":88},{"model":"Kimi K2 Instruct","score":0.3936,"effort":null,"percent":39.4,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":89},{"model":"Kimi K2 (Jul 2025)","score":0.3936,"effort":null,"percent":39.4,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":90},{"model":"GPT-4.1","score":0.3904,"effort":null,"percent":39.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":91,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.39,"effort":null,"percent":39.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":92,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":0.387,"effort":null,"percent":38.7,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":93},{"model":"Qwen3-235B-A22B-Thinking-2507","score":0.387,"effort":null,"percent":38.7,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":94},{"model":"DeepSeek-V3.1","score":0.3836,"effort":null,"percent":38.4,"date":"2025-08-21","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":95,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"GPT-5 nano","score":0.3806,"effort":"high","percent":38.1,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":96,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Nemotron 3 Super","score":0.3801,"effort":null,"percent":38.0,"date":"2026-03-11","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":97},{"model":"GPT-5.4 nano (no thinking)","score":0.3796,"effort":null,"percent":38.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":98},{"model":"GPT-4.1 mini","score":0.3761,"effort":null,"percent":37.6,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":99,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Qwen3-235B-A22B","score":0.3728,"effort":null,"percent":37.3,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":100}]},{"id":"mmlu","name":"MMLU","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2020-09-07","entrants":127,"leader":{"model":"GPT-4o (Nov 2024)","score":0.881,"effort":null,"percent":88.1,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1},"rows":[{"model":"GPT-4o (Nov 2024)","score":0.881,"effort":null,"percent":88.1,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.873,"effort":null,"percent":87.3,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"DeepSeek-V3","score":0.872,"effort":null,"percent":87.2,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":3,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.869,"effort":null,"percent":86.9,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.865,"effort":null,"percent":86.5,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"GPT-4 (Mar 2023)","score":0.864,"effort":null,"percent":86.4,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"Llama 3.3 70B","score":0.863,"effort":null,"percent":86.3,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":7},{"model":"Gemini 1.5 Pro (May 2024)","score":0.859,"effort":null,"percent":85.9,"date":"2024-05-14","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Qwen2.5-72B","score":0.853,"effort":null,"percent":85.3,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"Phi-4","score":0.848,"effort":null,"percent":84.8,"date":"2024-12-12","organization":"Microsoft Research","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":10},{"model":"Claude 3 Opus","score":0.846,"effort":null,"percent":84.6,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"Llama 3.1-405B","score":0.845,"effort":null,"percent":84.5,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":12},{"model":"GPT-4o (Aug 2024)","score":0.843,"effort":null,"percent":84.3,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"GPT-4o (May 2024)","score":0.842,"effort":null,"percent":84.2,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"Gemini 1.5 Pro (Feb 2024)","score":0.827,"effort":null,"percent":82.7,"date":"2024-02-15","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"GPT-4 (Jun 2023)","score":0.824,"effort":null,"percent":82.4,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"Qwen2-72B","score":0.824,"effort":null,"percent":82.4,"date":"2024-06-07","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17},{"model":"Amazon Nova Pro","score":0.82,"effort":null,"percent":82.0,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"GPT-4o mini","score":0.818,"effort":null,"percent":81.8,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-4 Turbo (Apr 2024)","score":0.813,"effort":null,"percent":81.3,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"Llama 3.2 90B","score":0.803,"effort":null,"percent":80.3,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"Llama 3.1-70B","score":0.801,"effort":null,"percent":80.1,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"Mistral Large 2 (Jul 2024)","score":0.8,"effort":null,"percent":80.0,"date":"2024-07-24","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":23},{"model":"Qwen2.5-14B","score":0.799,"effort":null,"percent":79.9,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":24},{"model":"Gemini 2.0 Flash (Dec 2024)","score":0.797,"effort":null,"percent":79.7,"date":"2024-12-11","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"GPT-4 Turbo (Nov 2023)","score":0.796,"effort":null,"percent":79.6,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26},{"model":"Llama 3-70B","score":0.793,"effort":null,"percent":79.3,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":27},{"model":"Yi-Large","score":0.793,"effort":null,"percent":79.3,"date":"2024-05-13","organization":"01.AI","countries":["China"],"accessibility":"API access","rank":28},{"model":"Qwen2.5-Coder-32B","score":0.791,"effort":null,"percent":79.1,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29},{"model":"Claude 2","score":0.785,"effort":null,"percent":78.5,"date":"2023-07-11","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.784,"effort":null,"percent":78.4,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":31},{"model":"phi-3-medium 14B","score":0.78,"effort":null,"percent":78.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"Gemini 1.5 Flash (May 2024)","score":0.779,"effort":null,"percent":77.9,"date":"2024-05-23","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":33},{"model":"Mixtral 8x22B","score":0.778,"effort":null,"percent":77.8,"date":"2024-04-17","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"Amazon Nova Lite","score":0.77,"effort":null,"percent":77.0,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":35},{"model":"Claude 1.3","score":0.77,"effort":null,"percent":77.0,"date":"2023-04-18","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"Yi-34B","score":0.763,"effort":null,"percent":76.3,"date":"2023-11-02","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":37},{"model":"Claude 3 Sonnet","score":0.759,"effort":null,"percent":75.9,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"Gemma 2 27B","score":0.757,"effort":null,"percent":75.7,"date":"2024-06-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":39},{"model":"phi-3-small 7.4B","score":0.757,"effort":null,"percent":75.7,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":40},{"model":"Qwen2.5-Coder-14B","score":0.752,"effort":null,"percent":75.2,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":41},{"model":"Qwen1.5-32B","score":0.744,"effort":null,"percent":74.4,"date":"2024-02-04","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":42},{"model":"Claude 3.5 Haiku (Oct 2024)","score":0.743,"effort":null,"percent":74.3,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":43},{"model":"Gemini 1.5 Flash (Sep 2024)","score":0.739,"effort":null,"percent":73.9,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"Claude 3 Haiku","score":0.738,"effort":null,"percent":73.8,"date":"2024-03-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":45},{"model":"Claude 2.1","score":0.735,"effort":null,"percent":73.5,"date":"2023-11-21","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":46},{"model":"Claude Instant","score":0.734,"effort":null,"percent":73.4,"date":null,"organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"Qwen2.5-7B","score":0.729,"effort":null,"percent":72.9,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":48},{"model":"Inflection-1","score":0.727,"effort":null,"percent":72.7,"date":"2023-06-22","organization":"Inflection AI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":49},{"model":"Gemma 2 9B","score":0.721,"effort":null,"percent":72.1,"date":"2024-06-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":50},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.714,"effort":null,"percent":71.4,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":51},{"model":"Amazon Nova Micro","score":0.708,"effort":null,"percent":70.8,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"Falcon-180B","score":0.706,"effort":null,"percent":70.6,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":53},{"model":"Mixtral 8x7B","score":0.706,"effort":null,"percent":70.6,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":54},{"model":"Gemini 1.0 Pro","score":0.7,"effort":null,"percent":70.0,"date":"2023-12-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":55},{"model":"davinci-002","score":0.7,"effort":null,"percent":70.0,"date":"2022-03-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"Llama 2-70B","score":0.699,"effort":null,"percent":69.9,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":57},{"model":"Command R+","score":0.694,"effort":null,"percent":69.4,"date":"2024-08-30","organization":"Cohere,Cohere Labs (formerly Cohere for AI)","countries":["Canada"],"accessibility":"Open weights (non-commercial)","rank":58},{"model":"PaLM (540B)","score":0.693,"effort":null,"percent":69.3,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":59},{"model":"GPT-3.5 Turbo (June 2023)","score":0.689,"effort":null,"percent":68.9,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":60},{"model":"Llama 3-8B","score":0.688,"effort":null,"percent":68.8,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":61},{"model":"Mistral Large","score":0.688,"effort":null,"percent":68.8,"date":"2024-02-26","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":62},{"model":"phi-3-mini 3.8B","score":0.688,"effort":null,"percent":68.8,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":63},{"model":"mistral-small-2402","score":0.687,"effort":null,"percent":68.7,"date":"2024-02-26","organization":null,"countries":[],"accessibility":null,"rank":64},{"model":"Qwen1.5-14B","score":0.686,"effort":null,"percent":68.6,"date":"2024-02-04","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":65},{"model":"Stable Beluga 2","score":0.686,"effort":null,"percent":68.6,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":66},{"model":"Yi-9B","score":0.684,"effort":null,"percent":68.4,"date":"2024-03-01","organization":null,"countries":[],"accessibility":null,"rank":67},{"model":"Qwen2.5-Coder (7B)","score":0.68,"effort":null,"percent":68.0,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":68},{"model":"Chinchilla","score":0.675,"effort":null,"percent":67.5,"date":"2022-03-29","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":69},{"model":"GPT-3.5 Turbo (Jan 2024)","score":0.673,"effort":null,"percent":67.3,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":70},{"model":"Qwen-14B","score":0.663,"effort":null,"percent":66.3,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":71},{"model":"Gemma 7B","score":0.661,"effort":null,"percent":66.1,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":72},{"model":"c4ai-command-r-08-2024","score":0.652,"effort":null,"percent":65.2,"date":"2024-08-30","organization":null,"countries":[],"accessibility":null,"rank":73},{"model":"StarCoder 2 15B","score":0.641,"effort":null,"percent":64.1,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":74},{"model":"Yi 6B","score":0.64,"effort":null,"percent":64.0,"date":"2023-11-02","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":75},{"model":"LLaMA-65B","score":0.634,"effort":null,"percent":63.4,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":76},{"model":"Llama 2-34B","score":0.626,"effort":null,"percent":62.6,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":77},{"model":"Qwen1.5-7B","score":0.626,"effort":null,"percent":62.6,"date":"2024-02-04","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":78},{"model":"Mistral 7B v0.2","score":0.625,"effort":null,"percent":62.5,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":79},{"model":"Mistral 7B v0.1","score":0.625,"effort":null,"percent":62.5,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":80},{"model":"DeepSeek-Coder-V2-Lite-Base","score":0.605,"effort":null,"percent":60.5,"date":"2024-06-13","organization":null,"countries":[],"accessibility":null,"rank":81},{"model":"Gopher (280B)","score":0.6,"effort":null,"percent":60.0,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":82},{"model":"Mistral 7B v0.3","score":0.599,"effort":null,"percent":59.9,"date":"2024-05-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":83},{"model":"Baichuan2-13B","score":0.592,"effort":null,"percent":59.2,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":84},{"model":"LLaMA-33B","score":0.587,"effort":null,"percent":58.7,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":85},{"model":"Nemotron-4 15B","score":0.587,"effort":null,"percent":58.7,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":86},{"model":"Falcon 2-11B","score":0.584,"effort":null,"percent":58.4,"date":"2024-05-09","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":87},{"model":"Phi-2","score":0.584,"effort":null,"percent":58.4,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":88},{"model":"internlm-chat-20b","score":0.574,"effort":null,"percent":57.4,"date":"2023-09-17","organization":null,"countries":[],"accessibility":null,"rank":89},{"model":"Falcon-40B","score":0.569,"effort":null,"percent":56.9,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":90},{"model":"Llama 3.2 11B","score":0.565,"effort":null,"percent":56.5,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":91},{"model":"Llama 3.1-8B","score":0.561,"effort":null,"percent":56.1,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":92},{"model":"Llama 2-13B","score":0.556,"effort":null,"percent":55.6,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":93},{"model":"Baichuan2-13B-Chat","score":0.551,"effort":null,"percent":55.1,"date":"2023-09-06","organization":null,"countries":[],"accessibility":null,"rank":94},{"model":"Baichuan 2-7B","score":0.5416,"effort":null,"percent":54.2,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":95},{"model":"PaLM 62B","score":0.537,"effort":null,"percent":53.7,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":96},{"model":"Qwen2.5-Coder (1.5B)","score":0.536,"effort":null,"percent":53.6,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":97},{"model":"Baichuan 1-13B","score":0.516,"effort":null,"percent":51.6,"date":"2023-07-11","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":98},{"model":"internlm-7b","score":0.51,"effort":null,"percent":51.0,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":99},{"model":"INTELLECT-1","score":0.4989,"effort":null,"percent":49.9,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":100}]},{"id":"lmca","name":"LMCA","in_index":true,"scale":0.01,"ceiling":0.85,"random_baseline":0.0,"released":"2026-08-12","entrants":123,"leader":{"model":"Claude Opus 5","score":63.3,"effort":"max","percent":null,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},"rows":[{"model":"Claude Opus 5","score":63.3,"effort":"max","percent":null,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Fable 5","score":60.34,"effort":"max","percent":null,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":59.21,"effort":null,"percent":null,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"GPT-5.6 Sol","score":58.43,"effort":"max","percent":null,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 4.8","score":57.51,"effort":null,"percent":null,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Claude Opus 4.6","score":55.76,"effort":"max","percent":null,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.5","score":54.27,"effort":"xhigh","percent":null,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.5 Pro","score":53.94,"effort":"xhigh","percent":null,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"Gemini 3.1 Pro Preview","score":53.82,"effort":null,"percent":null,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Kimi K3","score":52.74,"effort":"max","percent":null,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":10,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.6 Terra","score":52.23,"effort":"max","percent":null,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Opus 4.7","score":52.2,"effort":"max","percent":null,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.4","score":51.97,"effort":"xhigh","percent":null,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.7 Flash","score":50.38,"effort":"high","percent":null,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Muse Spark 1.1","score":49.91,"effort":"high","percent":null,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"Claude Sonnet 5","score":49.31,"effort":"max","percent":null,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"GPT-5.6 Luna","score":48.54,"effort":"max","percent":null,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Grok 4.6","score":48.45,"effort":"xhigh","percent":null,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Muse Spark 1.2","score":48.35,"effort":"xhigh","percent":null,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"Gemini 3.5 Flash","score":47.15,"effort":null,"percent":null,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Claude Sonnet 4.6","score":46.51,"effort":"max","percent":null,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Qwen3.8 Max","score":46.24,"effort":"xhigh","percent":null,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":22,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GLM-5.2","score":45.82,"effort":null,"percent":null,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Grok 4.5","score":45.2,"effort":"high","percent":null,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Gemini 3.6 Flash","score":44.93,"effort":null,"percent":null,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Claude Opus 4.5","score":44.46,"effort":"high","percent":null,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Qwen3.7 Max","score":44.02,"effort":null,"percent":null,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":27,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"GPT-5.2","score":43.91,"effort":"xhigh","percent":null,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"GPT-5.1","score":43.89,"effort":"high","percent":null,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Gemini 3 Flash Preview","score":43.13,"effort":null,"percent":null,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Qwen 3.6 Max (Preview)","score":42.53,"effort":null,"percent":null,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":31},{"model":"DeepSeek V4 Flash 0731","score":41.73,"effort":"max","percent":null,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":32,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"DeepSeek v4 Pro","score":41.2,"effort":"max","percent":null,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5.4 mini","score":40.79,"effort":"xhigh","percent":null,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GPT-5","score":39.99,"effort":"high","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o3","score":39.7,"effort":"high","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemma 4 31B IT","score":39.25,"effort":null,"percent":null,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":37,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"Claude Sonnet 4.5","score":38.76,"effort":"unknown thinking","percent":null,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Grok 4.20","score":38.66,"effort":null,"percent":null,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"o3-pro","score":38.48,"effort":null,"percent":null,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"Grok 4.3 Beta","score":38.33,"effort":null,"percent":null,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":41},{"model":"Qwen3.5 397B-A17B","score":37.89,"effort":null,"percent":null,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"Gemini 3.5 Flash-Lite","score":37.6,"effort":null,"percent":null,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Inkling","score":37.58,"effort":"xhigh","percent":null,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":44},{"model":"Qwen3.7 Plus","score":37.58,"effort":null,"percent":null,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":45,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Claude Opus 4","score":37.43,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Kimi K2.6","score":37.26,"effort":null,"percent":null,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Claude Opus 4.1","score":37.08,"effort":"unknown thinking","percent":null,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Nemotron 3 Ultra","score":36.9,"effort":null,"percent":null,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":49},{"model":"GPT-5.4 nano","score":36.86,"effort":"xhigh","percent":null,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Qwen3.5 Plus","score":36.36,"effort":null,"percent":null,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":51,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"DeepSeek v4 Flash","score":35.88,"effort":"max","percent":null,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":52,"model_id":"deepseek/deepseek-v4-flash","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash.toml"},{"model":"Gemini 3.1 Flash-Lite","score":35.01,"effort":null,"percent":null,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":53,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":34.77,"effort":null,"percent":null,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":54},{"model":"Qwen3.6 27B","score":34.53,"effort":null,"percent":null,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":55,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"GPT-5 mini","score":34.2,"effort":"high","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Qwen3.5 27B","score":33.98,"effort":null,"percent":null,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":57,"model_id":"alibaba/qwen3.5-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-27b.toml"},{"model":"MiniMax-M3","score":33.7,"effort":null,"percent":null,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":58},{"model":"Qwen 3.6 Plus (2026-04-02)","score":33.09,"effort":null,"percent":null,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":59},{"model":"Qwen3.5-122B-A10B","score":32.23,"effort":null,"percent":null,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":60,"model_id":"alibaba/qwen3.5-122b-a10b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-122b-a10b.toml"},{"model":"Qwen 3.6 Flash (2026-04-16)","score":30.98,"effort":null,"percent":null,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":61},{"model":"Claude Haiku 4.5","score":30.94,"effort":"unknown thinking","percent":null,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":62,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Qwen3.6 35B-A3B","score":29.7,"effort":null,"percent":null,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":63,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Gemma 4 26B A4B","score":29.68,"effort":null,"percent":null,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":64},{"model":"MiMo-V2.5-Pro","score":29.5,"effort":null,"percent":null,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":65},{"model":"Qwen3.5-35B-A3B","score":29.47,"effort":null,"percent":null,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":66,"model_id":"alibaba/qwen3.5-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-35b-a3b.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":29.31,"effort":null,"percent":null,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":67},{"model":"Qwen3.5 Flash","score":29.14,"effort":null,"percent":null,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":68,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"DeepSeek-V3.2-Exp","score":29.11,"effort":null,"percent":null,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":69},{"model":"Claude Sonnet 4","score":29.0,"effort":"unknown thinking","percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":70,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":28.82,"effort":null,"percent":null,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":71},{"model":"DeepSeek-V3.1-Terminus","score":28.58,"effort":null,"percent":null,"date":"2025-09-22","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":72},{"model":"Qwen3-Max-Instruct","score":28.27,"effort":null,"percent":null,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":73},{"model":"Gemini 2.5 Flash (Jun 2025)","score":27.49,"effort":null,"percent":null,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":74},{"model":"o4-mini","score":26.5,"effort":"high","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":75,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Mistral Medium 3.5","score":26.06,"effort":null,"percent":null,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":76,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"GPT-4.1","score":25.61,"effort":null,"percent":null,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Qwen3-235B-A22B","score":25.0,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":78},{"model":"Qwen3.5 9B","score":24.47,"effort":null,"percent":null,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":79,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"DeepSeek-V3.1","score":24.29,"effort":null,"percent":null,"date":"2025-08-21","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":80,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"Qwen Plus (Apr 2025)","score":23.96,"effort":null,"percent":null,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":81},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":23.52,"effort":null,"percent":null,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":82},{"model":"Qwen3-30B-A3B-Instruct (Jul 2025)","score":22.37,"effort":null,"percent":null,"date":"2025-07-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":83},{"model":"o1","score":22.27,"effort":"high","percent":null,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":84,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"gpt-oss-120b","score":22.14,"effort":null,"percent":null,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":85,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"GPT-4.1 mini","score":21.12,"effort":null,"percent":null,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":86,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Mistral Small 4","score":20.61,"effort":null,"percent":null,"date":"2026-03-16","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":87,"model_id":"mistral/mistral-small-2603","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-small-2603.toml"},{"model":"Qwen3-30B-A3B-Thinking (Jul 2025)","score":19.53,"effort":null,"percent":null,"date":"2025-07-30","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":88},{"model":"Mistral Medium 3.1","score":19.05,"effort":null,"percent":null,"date":"2025-08-12","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":89},{"model":"o3-mini","score":18.97,"effort":"high","percent":null,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":90,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Qwen3-14B","score":18.19,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":91},{"model":"gemini-2.5-flash-lite-preview-06-17 (Default thinking length)","score":18.07,"effort":null,"percent":null,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":92},{"model":"Llama 3.3 70B","score":17.52,"effort":null,"percent":null,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":93},{"model":"Qwen3-32B","score":17.28,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":94},{"model":"GPT-4 (Jun 2023)","score":17.06,"effort":null,"percent":null,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":95},{"model":"Claude 3 Opus","score":17.0,"effort":null,"percent":null,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":96},{"model":"Mistral Large 3","score":16.7,"effort":null,"percent":null,"date":"2025-12-02","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":97,"model_id":"mistral/mistral-large-2512","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-large-2512.toml"},{"model":"GPT-4o (May 2024)","score":16.59,"effort":null,"percent":null,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":98},{"model":"Llama 4 Maverick","score":15.87,"effort":null,"percent":null,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":99},{"model":"Qwen3-30B-A3B","score":15.82,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":100,"model_id":"alibaba/qwen3-30b-a3b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-30b-a3b.toml"}]},{"id":"math-level-5","name":"MATH level 5","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2021-03-05","entrants":98,"leader":{"model":"GPT-5","score":0.9813,"effort":"high","percent":98.1,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},"rows":[{"model":"GPT-5","score":0.9813,"effort":"high","percent":98.1,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5 mini","score":0.9785,"effort":"high","percent":97.8,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"o4-mini","score":0.9783,"effort":"high","percent":97.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"o3","score":0.9777,"effort":"high","percent":97.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.9773,"effort":null,"percent":97.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Qwen3-Max-Instruct","score":0.9713,"effort":null,"percent":97.1,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":6},{"model":"DeepSeek-R1 (May 2025)","score":0.9664,"effort":null,"percent":96.6,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"o3-mini","score":0.9649,"effort":"high","percent":96.5,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Claude Haiku 4.5","score":0.9636,"effort":null,"percent":96.4,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.959,"effort":null,"percent":95.9,"date":"2025-05-06","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.9556,"effort":null,"percent":95.6,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"GPT-5 nano","score":0.9524,"effort":"medium","percent":95.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"o1","score":0.9471,"effort":"high","percent":94.7,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"DeepSeek-R1","score":0.9305,"effort":null,"percent":93.1,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Claude 3.7 Sonnet (64k thinking)","score":0.9116,"effort":null,"percent":91.2,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Grok-3 mini","score":0.9094,"effort":null,"percent":90.9,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"Claude 3.7 Sonnet (32k thinking)","score":0.9003,"effort":null,"percent":90.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"DeepSeek-R1-Distill-Llama-70B","score":0.899,"effort":null,"percent":89.9,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18},{"model":"o1-mini","score":0.8918,"effort":"high","percent":89.2,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"Grok 3","score":0.8875,"effort":null,"percent":88.7,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"GPT-4.1 mini","score":0.8729,"effort":null,"percent":87.3,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"DeepSeek-R1-Distill-Qwen-14B","score":0.8712,"effort":null,"percent":87.1,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"Claude 3.7 Sonnet (16k thinking)","score":0.8625,"effort":null,"percent":86.3,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"Claude Opus 4","score":0.8505,"effort":null,"percent":85.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Claude Sonnet 4","score":0.8437,"effort":null,"percent":84.4,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Gemini 2.0 Pro Exp (Feb 2025)","score":0.8346,"effort":null,"percent":83.5,"date":"2025-02-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":26},{"model":"GPT-4.1","score":0.8301,"effort":null,"percent":83.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.8217,"effort":null,"percent":82.2,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"o1-preview","score":0.8165,"effort":null,"percent":81.6,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29},{"model":"Mistral Medium 3","score":0.8163,"effort":null,"percent":81.6,"date":"2025-05-07","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":30,"model_id":"mistral/mistral-medium-2505","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2505.toml"},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.7863,"effort":null,"percent":78.6,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"DeepSeek-V3 (Mar 2025)","score":0.7555,"effort":null,"percent":75.5,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":32},{"model":"Gemma 3 27B","score":0.7404,"effort":null,"percent":74.0,"date":"2025-03-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":33},{"model":"Llama 4 Maverick (FP8)","score":0.7302,"effort":null,"percent":73.0,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":34},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.7039,"effort":null,"percent":70.4,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":35},{"model":"GPT-4.1 nano","score":0.7,"effort":null,"percent":70.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"model":"Qwen3-235B-A22B","score":0.6886,"effort":null,"percent":68.9,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":37},{"model":"Claude 3.7 Sonnet","score":0.6818,"effort":null,"percent":68.2,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"Qwen2.5-Max","score":0.6718,"effort":null,"percent":67.2,"date":"2025-01-25","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":39},{"model":"Qwen Plus (Jan 2025)","score":0.6528,"effort":null,"percent":65.3,"date":"2025-01-25","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":40},{"model":"Phi-4","score":0.6494,"effort":null,"percent":64.9,"date":"2024-12-12","organization":"Microsoft Research","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"DeepSeek-V3","score":0.6485,"effort":null,"percent":64.9,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":42,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Grok-2 (Dec 2024)","score":0.6352,"effort":null,"percent":63.5,"date":"2024-12-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":43},{"model":"Qwen2.5-72B","score":0.6317,"effort":null,"percent":63.2,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44},{"model":"Llama 4 Scout","score":0.6227,"effort":null,"percent":62.3,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":45},{"model":"Gemini 1.5 Flash (Sep 2024)","score":0.6187,"effort":null,"percent":61.9,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":46},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.5695,"effort":null,"percent":56.9,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"Qwen-Turbo","score":0.5623,"effort":null,"percent":56.2,"date":"2024-11-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":48,"model_id":"alibaba/qwen-turbo","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen-turbo.toml"},{"model":"Qwen2.5-32B","score":0.5607,"effort":null,"percent":56.1,"date":"2024-09-17","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":49},{"model":"GPT-4o (Aug 2024)","score":0.5328,"effort":null,"percent":53.3,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":50},{"model":"GPT-4o mini","score":0.5263,"effort":null,"percent":52.6,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":51,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.5168,"effort":null,"percent":51.7,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"GPT-4o (May 2024)","score":0.5105,"effort":null,"percent":51.0,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"Mistral Large 2 (Nov 2024)","score":0.5028,"effort":null,"percent":50.3,"date":"2024-11-18","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":54},{"model":"GPT-4o (Nov 2024)","score":0.4977,"effort":null,"percent":49.8,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55},{"model":"Llama 3.1-405B","score":0.4977,"effort":null,"percent":49.8,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":56},{"model":"Mistral Small 3.1","score":0.4677,"effort":null,"percent":46.8,"date":"2025-03-17","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":57},{"model":"GPT-4 Turbo (Apr 2024)","score":0.4673,"effort":null,"percent":46.7,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":58},{"model":"Claude 3.5 Haiku (Oct 2024)","score":0.4636,"effort":null,"percent":46.4,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"Mistral Small 3","score":0.4482,"effort":null,"percent":44.8,"date":"2025-01-25","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":60},{"model":"Mistral Large 2 (Jul 2024)","score":0.4482,"effort":null,"percent":44.8,"date":"2024-07-24","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":61},{"model":"Tulu 3 (Tülu 3) 70B","score":0.4266,"effort":null,"percent":42.7,"date":"2024-11-21","organization":"Allen Institute for AI,University of Washington","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":62},{"model":"Llama 3.3 70B","score":0.416,"effort":null,"percent":41.6,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":63},{"model":"Gemini 1.5 Pro (May 2024)","score":0.4075,"effort":null,"percent":40.7,"date":"2024-05-14","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"GPT-4 Turbo Preview (Nov 2023)","score":0.4002,"effort":null,"percent":40.0,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":65},{"model":"Llama 3.2 90B","score":0.3944,"effort":null,"percent":39.4,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":66},{"model":"Qwen2-72B","score":0.3907,"effort":null,"percent":39.1,"date":"2024-06-07","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":67},{"model":"Claude 3 Opus","score":0.3748,"effort":null,"percent":37.5,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":68},{"model":"Llama 3.1-70B","score":0.3668,"effort":null,"percent":36.7,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":69},{"model":"GPT-4 Turbo Preview (January 2024)","score":0.3541,"effort":null,"percent":35.4,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":70},{"model":"Gemma 2 27B","score":0.2789,"effort":null,"percent":27.9,"date":"2024-06-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":71},{"model":"WizardLM-2 8x22B","score":0.2574,"effort":null,"percent":25.7,"date":"2024-04-15","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":72},{"model":"Yi-1.5-34B","score":0.2548,"effort":null,"percent":25.5,"date":"2024-05-13","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":73},{"model":"Gemini 1.5 Flash (May 2024)","score":0.2512,"effort":null,"percent":25.1,"date":"2024-05-23","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":74},{"model":"Mistral Large","score":0.2446,"effort":null,"percent":24.5,"date":"2024-02-26","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":75},{"model":"Mixtral 8x22B","score":0.2424,"effort":null,"percent":24.2,"date":"2024-04-17","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":76},{"model":"GPT-4 (Jun 2023)","score":0.2297,"effort":null,"percent":23.0,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"Llama 3.1-8B","score":0.2288,"effort":null,"percent":22.9,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":78},{"model":"Hermes 2 Theta Llama-3 70B","score":0.2269,"effort":null,"percent":22.7,"date":"2024-06-20","organization":"Nous Research,Arcee AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":79},{"model":"Llama 3-70B","score":0.2255,"effort":null,"percent":22.6,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":80},{"model":"Gemma 2 9B","score":0.2101,"effort":null,"percent":21.0,"date":"2024-06-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":81},{"model":"Claude 3 Sonnet","score":0.1817,"effort":null,"percent":18.2,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":82},{"model":"phi-3-medium 14B","score":0.1756,"effort":null,"percent":17.6,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":83},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.1589,"effort":null,"percent":15.9,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":84},{"model":"Ministral 8B","score":0.1494,"effort":null,"percent":14.9,"date":"2024-10-16","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":85},{"model":"Claude 3 Haiku","score":0.1488,"effort":null,"percent":14.9,"date":"2024-03-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":86},{"model":"Ministral 3B","score":0.1444,"effort":null,"percent":14.4,"date":"2024-10-16","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":87,"model_id":"mistral/ministral-3b","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/ministral-3b.toml"},{"model":"Claude 2","score":0.1173,"effort":null,"percent":11.7,"date":"2023-07-11","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":88},{"model":"DBRX","score":0.1165,"effort":null,"percent":11.7,"date":"2024-03-27","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":89},{"model":"GPT-3.5 Turbo (Jan 2024)","score":0.1163,"effort":null,"percent":11.6,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":90},{"model":"Gemini 1.0 Pro","score":0.1124,"effort":null,"percent":11.2,"date":"2023-12-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":91},{"model":"Mistral NeMo","score":0.1083,"effort":null,"percent":10.8,"date":"2024-07-18","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":92,"model_id":"mistral/mistral-nemo","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-nemo.toml"},{"model":"Mixtral 8x7B","score":0.0995,"effort":null,"percent":10.0,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":93},{"model":"DeepSeek LLM 67B","score":0.0639,"effort":null,"percent":6.4,"date":"2023-11-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":94},{"model":"Llama 3-8B","score":0.0613,"effort":null,"percent":6.1,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":95},{"model":"Yi-34B","score":0.0515,"effort":null,"percent":5.1,"date":"2023-11-22","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":96},{"model":"Mistral 7B v0.3","score":0.0368,"effort":null,"percent":3.7,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":97},{"model":"Llama 2-70B","score":0.0329,"effort":null,"percent":3.3,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":98}]},{"id":"arc-agi","name":"ARC-AGI","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2019-11-05","entrants":95,"leader":{"model":"Claude Fable 5","score":0.985,"effort":"xhigh","percent":98.5,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.985,"effort":"xhigh","percent":98.5,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-6 Astra","score":0.985,"effort":"xhigh","percent":98.5,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.98,"effort":null,"percent":98.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.6 Sol","score":0.975,"effort":"xhigh","percent":97.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 5","score":0.975,"effort":"max","percent":97.5,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Fable 5.1","score":0.975,"effort":"max","percent":97.5,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-5.5 Pro","score":0.965,"effort":"high","percent":96.5,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"GPT-5.6 Terra","score":0.965,"effort":"max","percent":96.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Gemini 3 Deep Think","score":0.96,"effort":null,"percent":96.0,"date":"2026-02-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":9},{"model":"Gemini 3.7 Flash","score":0.955,"effort":"high","percent":95.5,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GPT-5.5","score":0.95,"effort":"xhigh","percent":95.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4 Pro","score":0.945,"effort":"xhigh","percent":94.5,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Kimi K3","score":0.945,"effort":"max","percent":94.5,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":13,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Claude Opus 4.6 (120k thinking)","score":0.94,"effort":null,"percent":94.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"GPT-5.4","score":0.9367,"effort":"xhigh","percent":93.7,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Claude Opus 4.7","score":0.935,"effort":"high","percent":93.5,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Gemini 3.5 Flash","score":0.925,"effort":null,"percent":92.5,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Claude Opus 4.8","score":0.925,"effort":null,"percent":92.5,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Gemini 3.6 Flash","score":0.9117,"effort":null,"percent":91.2,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GPT-5.2 Pro","score":0.905,"effort":null,"percent":90.5,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.905,"effort":"low","percent":90.5,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Grok 4.20","score":0.895,"effort":null,"percent":89.5,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":22},{"model":"DeepSeek V4 Flash 0731","score":0.89,"effort":"max","percent":89.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"GPT-5.6 Luna","score":0.88,"effort":"max","percent":88.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Grok 4.6","score":0.875,"effort":"medium","percent":87.5,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Grok 4.5","score":0.8717,"effort":"medium","percent":87.2,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Claude Sonnet 4.6","score":0.865,"effort":"high","percent":86.5,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"GPT-5.2","score":0.862,"effort":"xhigh","percent":86.2,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3 Flash Preview","score":0.8467,"effort":null,"percent":84.7,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Inkling Small","score":0.84,"effort":"xhigh","percent":84.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":30},{"model":"Claude Opus 4.5 (64k thinking)","score":0.8,"effort":null,"percent":80.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"Inkling","score":0.795,"effort":null,"percent":79.5,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"GLM-5.2","score":0.77,"effort":null,"percent":77.0,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.7583,"effort":null,"percent":75.8,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34},{"model":"Gemini 3 Pro Preview","score":0.75,"effort":null,"percent":75.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.1","score":0.7283,"effort":"high","percent":72.8,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Claude Opus 4.5 (16k thinking)","score":0.72,"effort":null,"percent":72.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"GPT-5 Pro","score":0.702,"effort":null,"percent":70.2,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"Grok 4","score":0.6667,"effort":null,"percent":66.7,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"GPT-5","score":0.6567,"effort":"high","percent":65.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Kimi K2.5","score":0.6533,"effort":null,"percent":65.3,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.6367,"effort":null,"percent":63.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":42},{"model":"MiniMax-M2.5","score":0.6367,"effort":null,"percent":63.7,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43},{"model":"GPT-5.4 mini","score":0.6367,"effort":"xhigh","percent":63.7,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":44,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"o3","score":0.6083,"effort":"high","percent":60.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"o3-pro","score":0.5933,"effort":null,"percent":59.3,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"o4-mini","score":0.5867,"effort":"high","percent":58.7,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":47,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Opus 4.5 (8k thinking)","score":0.5867,"effort":null,"percent":58.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.57,"effort":null,"percent":57.0,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":49},{"model":"GPT-5 mini","score":0.5433,"effort":"high","percent":54.3,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.535,"effort":null,"percent":53.5,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":51,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"GPT-5.4 nano","score":0.515,"effort":"xhigh","percent":51.5,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Grok 4 Fast","score":0.485,"effort":null,"percent":48.5,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.4833,"effort":null,"percent":48.3,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":54},{"model":"Claude Haiku 4.5","score":0.4767,"effort":null,"percent":47.7,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Claude Sonnet 4.5 (8k thinking)","score":0.465,"effort":null,"percent":46.5,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"GLM-5","score":0.4467,"effort":null,"percent":44.7,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":57,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.41,"effort":null,"percent":41.0,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":58},{"model":"Claude Sonnet 4","score":0.4,"effort":null,"percent":40.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":59,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"tiny-recursion-model","score":0.4,"effort":null,"percent":40.0,"date":"2025-10-06","organization":null,"countries":[],"accessibility":null,"rank":60},{"model":"Claude Opus 4.5 (no thinking)","score":0.4,"effort":null,"percent":40.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":61},{"model":"Claude Opus 4","score":0.3567,"effort":null,"percent":35.7,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":62,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"o3-mini","score":0.345,"effort":"high","percent":34.5,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":63,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":0.3333,"effort":null,"percent":33.3,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.33,"effort":null,"percent":33.0,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":65},{"model":"Gemini 2.5 Flash (Apr 2025)","score":0.323,"effort":null,"percent":32.3,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":66},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.313,"effort":null,"percent":31.3,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":67},{"model":"Claude Sonnet 4.5 (1k thinking)","score":0.31,"effort":null,"percent":31.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":68},{"model":"o1","score":0.307,"effort":"medium","percent":30.7,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":69,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude 3.7 Sonnet (16k thinking)","score":0.286,"effort":null,"percent":28.6,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":70},{"model":"codex-mini-2025-05-16","score":0.2733,"effort":null,"percent":27.3,"date":"2025-05-16","organization":null,"countries":[],"accessibility":null,"rank":71},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.255,"effort":null,"percent":25.5,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":72},{"model":"o1 Pro","score":0.233,"effort":"low","percent":23.3,"date":"2025-03-19","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73,"model_id":"openai/o1-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1-pro.toml"},{"model":"DeepSeek-R1 (May 2025)","score":0.2121,"effort":null,"percent":21.2,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":74},{"model":"Claude 3.7 Sonnet (8k thinking)","score":0.212,"effort":null,"percent":21.2,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":75},{"model":"GPT-5 nano","score":0.2071,"effort":"medium","percent":20.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":76,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"o1-preview","score":0.18,"effort":null,"percent":18.0,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"Grok-3 mini","score":0.165,"effort":null,"percent":16.5,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":78},{"model":"Grok 3 mini","score":0.165,"effort":null,"percent":16.5,"date":"2025-06-24","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":79},{"model":"DeepSeek-R1","score":0.158,"effort":null,"percent":15.8,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":80,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"o1-mini","score":0.14,"effort":"medium","percent":14.0,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":81},{"model":"Claude 3.7 Sonnet","score":0.136,"effort":null,"percent":13.6,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":82},{"model":"Claude 3.7 Sonnet (1k thinking)","score":0.116,"effort":null,"percent":11.6,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":83},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":0.11,"effort":null,"percent":11.0,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":84},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.103,"effort":null,"percent":10.3,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":85},{"model":"Magistral Medium 1.1","score":0.0612,"effort":null,"percent":6.1,"date":"2025-06-10","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":86},{"model":"GPT-5.1 (no thinking)","score":0.0583,"effort":null,"percent":5.8,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":87},{"model":"GPT-4.1","score":0.055,"effort":null,"percent":5.5,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":88,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Grok 3","score":0.055,"effort":null,"percent":5.5,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":89},{"model":"Magistral Small 1.0","score":0.05,"effort":null,"percent":5.0,"date":"2025-06-10","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":90},{"model":"GPT-4o (Nov 2024)","score":0.045,"effort":null,"percent":4.5,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":91},{"model":"Llama 4 Maverick","score":0.0438,"effort":null,"percent":4.4,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":92},{"model":"GPT-4.1 mini","score":0.035,"effort":null,"percent":3.5,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":93,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Llama 4 Scout","score":0.005,"effort":null,"percent":0.5,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":94},{"model":"GPT-4.1 nano","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":95,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"}]},{"id":"arc-agi-2","name":"ARC-AGI-2","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-03-24","entrants":91,"leader":{"model":"GPT-6 Astra","score":0.95,"effort":"max","percent":95.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.95,"effort":"max","percent":95.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"GPT-5.6 Sol","score":0.925,"effort":"max","percent":92.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 5","score":0.9042,"effort":"max","percent":90.4,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Fable 5.1","score":0.9,"effort":"max","percent":90.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Fable 5","score":0.8917,"effort":"max","percent":89.2,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.5","score":0.85,"effort":"xhigh","percent":85.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.7 Flash","score":0.8458,"effort":"high","percent":84.6,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Gemini 3 Deep Think","score":0.8458,"effort":null,"percent":84.6,"date":"2026-02-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":8},{"model":"GPT-5.5 Pro","score":0.8458,"effort":"high","percent":84.6,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"GPT-5.6 Terra","score":0.839,"effort":"max","percent":83.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.4 Pro","score":0.8333,"effort":"xhigh","percent":83.3,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.771,"effort":null,"percent":77.1,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Opus 4.7","score":0.7583,"effort":"max","percent":75.8,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.4","score":0.7395,"effort":"xhigh","percent":74.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.5 Flash","score":0.7208,"effort":null,"percent":72.1,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Claude Opus 4.8","score":0.7208,"effort":null,"percent":72.1,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Claude Opus 4.6 (120k thinking)","score":0.6917,"effort":null,"percent":69.2,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"Grok 4.6","score":0.6708,"effort":"xhigh","percent":67.1,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Grok 4.20","score":0.6514,"effort":null,"percent":65.1,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"DeepSeek V4 Flash 0731","score":0.6139,"effort":"max","percent":61.4,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":20,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.6125,"effort":"max","percent":61.3,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Claude Sonnet 4.6","score":0.6042,"effort":"high","percent":60.4,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Kimi K3","score":0.6042,"effort":"max","percent":60.4,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":23,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.6 Flash","score":0.6042,"effort":null,"percent":60.4,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GPT-5.6 Luna","score":0.5954,"effort":"max","percent":59.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"GPT-5.2 Pro","score":0.5416,"effort":null,"percent":54.2,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"GPT-5.2","score":0.5291,"effort":"xhigh","percent":52.9,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Grok 4.5","score":0.5264,"effort":"high","percent":52.6,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Inkling Small","score":0.4014,"effort":"xhigh","percent":40.1,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":29},{"model":"Claude Opus 4.5 (64k thinking)","score":0.3764,"effort":null,"percent":37.6,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"Inkling","score":0.3653,"effort":null,"percent":36.5,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":31},{"model":"Gemini 3 Flash Preview","score":0.3361,"effort":null,"percent":33.6,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Gemini 3 Pro Preview","score":0.3111,"effort":null,"percent":31.1,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Claude Opus 4.5 (16k thinking)","score":0.2278,"effort":null,"percent":22.8,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34},{"model":"GLM-5.2","score":0.2278,"effort":null,"percent":22.8,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":35,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"GPT-5.4 mini","score":0.189,"effort":"xhigh","percent":18.9,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GPT-5 Pro","score":0.1833,"effort":null,"percent":18.3,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"GPT-5.1","score":0.1764,"effort":"high","percent":17.6,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Grok 4","score":0.1598,"effort":null,"percent":16.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"Claude Opus 4.5 (8k thinking)","score":0.1389,"effort":null,"percent":13.9,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.1361,"effort":null,"percent":13.6,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"Kimi K2.5","score":0.1181,"effort":null,"percent":11.8,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42},{"model":"Gemini 3.5 Flash-Lite","score":0.1028,"effort":null,"percent":10.3,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"GPT-5","score":0.0986,"effort":"high","percent":9.9,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":44,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Opus 4","score":0.0861,"effort":null,"percent":8.6,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.0778,"effort":null,"percent":7.8,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":46},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.0694,"effort":null,"percent":6.9,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"Claude Sonnet 4.5 (8k thinking)","score":0.0694,"effort":null,"percent":6.9,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"o3","score":0.0653,"effort":"high","percent":6.5,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"tiny-recursion-model","score":0.0625,"effort":null,"percent":6.2,"date":"2025-10-06","organization":null,"countries":[],"accessibility":null,"rank":50},{"model":"o4-mini","score":0.0611,"effort":"high","percent":6.1,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":51,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Sonnet 4","score":0.0593,"effort":null,"percent":5.9,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude Sonnet 4.5 (1k thinking)","score":0.0583,"effort":null,"percent":5.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"GPT-5.4 nano","score":0.0569,"effort":"xhigh","percent":5.7,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":54,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Grok 4 Fast","score":0.0528,"effort":null,"percent":5.3,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":55},{"model":"o3-pro","score":0.0486,"effort":null,"percent":4.9,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.0486,"effort":null,"percent":4.9,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":57},{"model":"MiniMax-M2.5","score":0.0486,"effort":null,"percent":4.9,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":58},{"model":"GLM-5","score":0.0486,"effort":null,"percent":4.9,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":59,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5 mini","score":0.0444,"effort":"high","percent":4.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":60,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Claude Haiku 4.5","score":0.0403,"effort":null,"percent":4.0,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":61,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.0403,"effort":null,"percent":4.0,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":62},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.0375,"effort":null,"percent":3.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":63},{"model":"o3-mini","score":0.0299,"effort":"high","percent":3.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":64,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"GPT-5 nano","score":0.0261,"effort":"high","percent":2.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":65,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":0.0254,"effort":null,"percent":2.5,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":66},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.013,"effort":null,"percent":1.3,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":67},{"model":"DeepSeek-R1","score":0.013,"effort":null,"percent":1.3,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":68,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"codex-mini-2025-05-16","score":0.0127,"effort":null,"percent":1.3,"date":"2025-05-16","organization":null,"countries":[],"accessibility":null,"rank":69},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":0.0125,"effort":null,"percent":1.2,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":70},{"model":"DeepSeek-R1 (May 2025)","score":0.0112,"effort":null,"percent":1.1,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":71},{"model":"Claude 3.7 Sonnet (8k thinking)","score":0.009,"effort":null,"percent":0.9,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":72},{"model":"o1-mini","score":0.0083,"effort":"unknown thinking","percent":0.8,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.008,"effort":null,"percent":0.8,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":74},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.008,"effort":null,"percent":0.8,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":75},{"model":"Claude 3.7 Sonnet (16k thinking)","score":0.007,"effort":null,"percent":0.7,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":76},{"model":"GPT-4.1","score":0.0042,"effort":null,"percent":0.4,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Grok 3 mini","score":0.0042,"effort":null,"percent":0.4,"date":"2025-06-24","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":78},{"model":"GPT-5.1 (no thinking)","score":0.0042,"effort":null,"percent":0.4,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":79},{"model":"Claude 3.7 Sonnet (1k thinking)","score":0.004,"effort":null,"percent":0.4,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":80},{"model":"Claude 3.7 Sonnet","score":0.0,"effort":null,"percent":0.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":81},{"model":"GPT-4o mini","score":0.0,"effort":null,"percent":0.0,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":82,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-4.1 nano","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":83,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"model":"Llama 4 Scout","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":84},{"model":"Llama 4 Maverick","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":85},{"model":"GPT-4.1 mini","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":86,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Grok 3","score":0.0,"effort":null,"percent":0.0,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":87},{"model":"Magistral Small 1.0","score":0.0,"effort":null,"percent":0.0,"date":"2025-06-10","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":88},{"model":"Magistral Medium 1.1","score":0.0,"effort":null,"percent":0.0,"date":"2025-06-10","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":89},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.0,"effort":null,"percent":0.0,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":90},{"model":"GPT-4o (Nov 2024)","score":0.0,"effort":null,"percent":0.0,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":91}]},{"id":"simplebench","name":"SimpleBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.1666666666666666,"released":"2024-10-31","entrants":89,"leader":{"model":"Claude Fable 5","score":0.819,"effort":"max","percent":81.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.819,"effort":"max","percent":81.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":0.806,"effort":"unknown","percent":80.6,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.796,"effort":null,"percent":79.6,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.5 Pro","score":0.769,"effort":"high","percent":76.9,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"Gemini 3.5 Flash","score":0.767,"effort":null,"percent":76.7,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Gemini 3 Pro Preview","score":0.764,"effort":null,"percent":76.4,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Grok 4.6","score":0.759,"effort":"unknown","percent":75.9,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Muse Spark 1.2","score":0.745,"effort":null,"percent":74.5,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"GPT-5.4 Pro","score":0.741,"effort":null,"percent":74.1,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"GPT-5.6 Sol (pro, xhigh)","score":0.717,"effort":null,"percent":71.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"GPT-5.6 Sol (pro, unknown thinking)","score":0.717,"effort":null,"percent":71.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"Qwen3.7 Max","score":0.704,"effort":null,"percent":70.4,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":12,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Grok 4.5","score":0.7,"effort":"unknown thinking","percent":70.0,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"GPT-5.5","score":0.69,"effort":"unknown thinking","percent":69.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.676,"effort":null,"percent":67.6,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Claude Opus 4.8","score":0.648,"effort":null,"percent":64.8,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.6 Sol","score":0.648,"effort":"unknown thinking","percent":64.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Qwen 3.6 Max (Preview)","score":0.63,"effort":null,"percent":63.0,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":18},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.624,"effort":null,"percent":62.4,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"Claude Opus 4.5 (no thinking)","score":0.62,"effort":null,"percent":62.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"Claude Opus 4.5","score":0.62,"effort":"unknown thinking","percent":62.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Claude Opus 4.7","score":0.617,"effort":"unknown","percent":61.7,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5 Pro","score":0.616,"effort":null,"percent":61.6,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"Gemini 3 Flash Preview","score":0.611,"effort":null,"percent":61.1,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.611,"effort":"unknown","percent":61.1,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":25,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Kimi K3","score":0.607,"effort":"max","percent":60.7,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":26,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Claude Sonnet 5","score":0.606,"effort":"unknown thinking","percent":60.6,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Grok 4","score":0.605,"effort":null,"percent":60.5,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Claude Opus 4.1","score":0.6,"effort":null,"percent":60.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Claude Opus 4","score":0.588,"effort":null,"percent":58.8,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"GLM-5.2","score":0.588,"effort":null,"percent":58.8,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":31,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Kimi K2.7 Code","score":0.579,"effort":null,"percent":57.9,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":32,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5.2 Pro","score":0.574,"effort":null,"percent":57.4,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"GPT-5","score":0.567,"effort":"high","percent":56.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Grok 4.1 Fast","score":0.56,"effort":null,"percent":56.0,"date":"2025-11-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"xai/grok-4.1-fast","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.1-fast.toml"},{"model":"GLM-5.1","score":0.551,"effort":null,"percent":55.1,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":36,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Claude Sonnet 4.5 (12k thinking)","score":0.543,"effort":null,"percent":54.3,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"Claude Sonnet 4.5","score":0.543,"effort":"unknown thinking","percent":54.3,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"GPT-5.1","score":0.532,"effort":"high","percent":53.2,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"GLM-5","score":0.532,"effort":null,"percent":53.2,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":40,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"o3","score":0.531,"effort":"high","percent":53.1,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"DeepSeek-V3.2-Speciale","score":0.526,"effort":null,"percent":52.6,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.516,"effort":null,"percent":51.6,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":43},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.516,"effort":null,"percent":51.6,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"Inkling","score":0.5,"effort":null,"percent":50.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":45},{"model":"GPT-5.6 Terra","score":0.489,"effort":"unknown thinking","percent":48.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GLM-4.7 (Novita)","score":0.477,"effort":null,"percent":47.7,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47},{"model":"GLM-4.7","score":0.477,"effort":null,"percent":47.7,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":48,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"GPT-5.6 Luna","score":0.468,"effort":"unknown thinking","percent":46.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Kimi K2.5","score":0.468,"effort":null,"percent":46.8,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":50},{"model":"Claude 3.7 Sonnet (12k thinking)","score":0.464,"effort":null,"percent":46.4,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":51},{"model":"Claude 3.7 Sonnet","score":0.464,"effort":"unknown thinking","percent":46.4,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"DeepSeek v4","score":0.463,"effort":"unknown","percent":46.3,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":53},{"model":"GPT-5.2","score":0.458,"effort":"high","percent":45.8,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":54,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"MiniMax-M3","score":0.458,"effort":null,"percent":45.8,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":55},{"model":"Claude Sonnet 4","score":0.455,"effort":null,"percent":45.5,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":56,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"o1-preview","score":0.417,"effort":null,"percent":41.7,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":57},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.414,"effort":null,"percent":41.4,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":58},{"model":"Gemini 2.5 Flash (Jun 2025)","score":0.412,"effort":null,"percent":41.2,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"DeepSeek-R1 (May 2025)","score":0.408,"effort":null,"percent":40.8,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":60},{"model":"o1","score":0.401,"effort":"high","percent":40.1,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":61,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"DeepSeek-V3.1","score":0.4,"effort":null,"percent":40.0,"date":"2025-08-21","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":62,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"o4-mini","score":0.387,"effort":"high","percent":38.7,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":63,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Grok 3","score":0.361,"effort":null,"percent":36.1,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":64},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.352,"effort":null,"percent":35.2,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":65},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.345,"effort":null,"percent":34.5,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":66},{"model":"Gemini 2.0 Flash (Dec 2024)","score":0.311,"effort":null,"percent":31.1,"date":"2024-12-06","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":67},{"model":"Qwen3-235B-A22B","score":0.31,"effort":null,"percent":31.0,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":68},{"model":"DeepSeek-R1","score":0.309,"effort":null,"percent":30.9,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":69,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Gemini 2.0 Flash Thinking Exp","score":0.307,"effort":null,"percent":30.7,"date":"2025-01-21","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":70},{"model":"Llama 4 Maverick","score":0.277,"effort":null,"percent":27.7,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":71},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.275,"effort":null,"percent":27.5,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":72},{"model":"DeepSeek-V3 (Mar 2025)","score":0.272,"effort":null,"percent":27.2,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":73},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.271,"effort":null,"percent":27.1,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":74},{"model":"GPT-4.1","score":0.27,"effort":null,"percent":27.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":75,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Kimi K2 Instruct","score":0.263,"effort":null,"percent":26.3,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":76},{"model":"GPT-4 Turbo (Apr 2024)","score":0.251,"effort":null,"percent":25.1,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":77},{"model":"Claude 3 Opus","score":0.235,"effort":null,"percent":23.5,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":78},{"model":"Llama 3.1-405B","score":0.23,"effort":null,"percent":23.0,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":79},{"model":"o3-mini","score":0.228,"effort":"high","percent":22.8,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":80,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Grok-2 (Dec 2024)","score":0.227,"effort":null,"percent":22.7,"date":"2024-12-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":81},{"model":"Mistral Large 2 (Jul 2024)","score":0.225,"effort":null,"percent":22.5,"date":"2024-07-24","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":82},{"model":"gpt-oss-120b","score":0.221,"effort":null,"percent":22.1,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":83,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Llama 3.3 70B","score":0.199,"effort":null,"percent":19.9,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":84},{"model":"DeepSeek-V3","score":0.189,"effort":null,"percent":18.9,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":85,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"o1-mini","score":0.181,"effort":"medium","percent":18.1,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":86},{"model":"GPT-4o (Aug 2024)","score":0.178,"effort":null,"percent":17.8,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":87},{"model":"Command R+","score":0.174,"effort":null,"percent":17.4,"date":"2024-08-30","organization":"Cohere,Cohere Labs (formerly Cohere for AI)","countries":["Canada"],"accessibility":"Open weights (non-commercial)","rank":88},{"model":"GPT-4o mini","score":0.107,"effort":null,"percent":10.7,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":89,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"}]},{"id":"gsm8k","name":"GSM8K","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2021-10-27","entrants":82,"leader":{"model":"DeepSeek-Coder-V2 236B","score":0.945,"effort":null,"percent":94.5,"date":"2024-06-17","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":1},"rows":[{"model":"DeepSeek-Coder-V2 236B","score":0.945,"effort":null,"percent":94.5,"date":"2024-06-17","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":1},{"model":"Qwen2.5-Coder-14B-Instruct","score":0.942,"effort":null,"percent":94.2,"date":"2024-11-06","organization":null,"countries":[],"accessibility":null,"rank":2},{"model":"Qwen2.5-Coder-32B-Instruct","score":0.93,"effort":null,"percent":93.0,"date":"2024-11-21","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":3,"model_id":"alibaba/qwen2.5-coder-32b-instruct","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-32b-instruct.toml"},{"model":"GPT-4 (Mar 2023)","score":0.92,"effort":null,"percent":92.0,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"GPT-4o mini","score":0.913,"effort":null,"percent":91.3,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Qwen2.5-Coder-32B","score":0.911,"effort":null,"percent":91.1,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6},{"model":"GPT-4 (Jun 2023)","score":0.8999,"effort":null,"percent":90.0,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Phi-3.5-MoE","score":0.887,"effort":null,"percent":88.7,"date":"2024-08-17","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":8},{"model":"Qwen2.5-Coder-14B","score":0.887,"effort":null,"percent":88.7,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":9},{"model":"DeepSeek-Coder-V2-Lite-Instruct","score":0.876,"effort":null,"percent":87.6,"date":"2024-06-13","organization":null,"countries":[],"accessibility":null,"rank":10},{"model":"Qwen2.5-Coder (7B)","score":0.867,"effort":null,"percent":86.7,"date":"2024-09-17","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"Claude Instant","score":0.867,"effort":null,"percent":86.7,"date":"2023-08-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Phi-3.5-mini","score":0.862,"effort":null,"percent":86.2,"date":"2024-08-16","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":13},{"model":"Gemma 2 9B","score":0.849,"effort":null,"percent":84.9,"date":null,"organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":14},{"model":"Mistral NeMo","score":0.842,"effort":null,"percent":84.2,"date":"2024-07-18","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":15,"model_id":"mistral/mistral-nemo","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-nemo.toml"},{"model":"Llama 3.1-8B","score":0.824,"effort":null,"percent":82.4,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":16},{"model":"Gemini 1.5 Flash (May 2024)","score":0.824,"effort":null,"percent":82.4,"date":"2024-05-23","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"Qwen2.5-Coder-3B-Instruct","score":0.807,"effort":null,"percent":80.7,"date":"2024-11-06","organization":null,"countries":[],"accessibility":null,"rank":18},{"model":"davinci-003","score":0.782,"effort":null,"percent":78.2,"date":"2022-11-28","organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":19},{"model":"Yi-34B","score":0.76,"effort":null,"percent":76.0,"date":"2023-11-22","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Qwen2.5-Coder-3B","score":0.757,"effort":null,"percent":75.7,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":21},{"model":"Mixtral 8x7B","score":0.744,"effort":null,"percent":74.4,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"Llama 2-70B","score":0.696,"effort":null,"percent":69.6,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":23},{"model":"Stable Beluga 2","score":0.696,"effort":null,"percent":69.6,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":24},{"model":"DeepSeek-Coder-V2-Lite-Base","score":0.671,"effort":null,"percent":67.1,"date":"2024-06-13","organization":null,"countries":[],"accessibility":null,"rank":25},{"model":"Qwen2.5-Coder (1.5B)","score":0.658,"effort":null,"percent":65.8,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26},{"model":"internlm-20b","score":0.629,"effort":null,"percent":62.9,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":27},{"model":"Qwen-14B","score":0.613,"effort":null,"percent":61.3,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":28},{"model":"GPT-3.5 Turbo (June 2023)","score":0.5777,"effort":null,"percent":57.8,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29},{"model":"StarCoder 2 15B","score":0.577,"effort":null,"percent":57.7,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":30},{"model":"davinci-002","score":0.568,"effort":null,"percent":56.8,"date":"2022-03-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"PaLM (540B)","score":0.565,"effort":null,"percent":56.5,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":32},{"model":"LLaMA-65B","score":0.544,"effort":null,"percent":54.4,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":33},{"model":"Mistral 7B v0.1","score":0.544,"effort":null,"percent":54.4,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"Falcon-180B","score":0.544,"effort":null,"percent":54.4,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":35},{"model":"Falcon 2-11B","score":0.5383,"effort":null,"percent":53.8,"date":"2024-05-09","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":36},{"model":"Baichuan2-13B","score":0.528,"effort":null,"percent":52.8,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":37},{"model":"Qwen-7B","score":0.517,"effort":null,"percent":51.7,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":38},{"model":"Gemma 7B","score":0.464,"effort":null,"percent":46.4,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":39},{"model":"Nemotron-4 15B","score":0.46,"effort":null,"percent":46.0,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":40},{"model":"Baichuan2-13B-Chat","score":0.457,"effort":null,"percent":45.7,"date":"2023-09-06","organization":null,"countries":[],"accessibility":null,"rank":41},{"model":"Yi 6B","score":0.449,"effort":null,"percent":44.9,"date":"2023-11-22","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":42},{"model":"LLaMA-33B","score":0.441,"effort":null,"percent":44.1,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":43},{"model":"Llama 2-34B","score":0.422,"effort":null,"percent":42.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":44},{"model":"INTELLECT-1","score":0.3858,"effort":null,"percent":38.6,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":45},{"model":"CodeQwen1.5-7B","score":0.377,"effort":null,"percent":37.7,"date":"2024-04-15","organization":null,"countries":[],"accessibility":null,"rank":46},{"model":"Llama 2-13B","score":0.369,"effort":null,"percent":36.9,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":47},{"model":"Mistral 7B v0.2","score":0.354,"effort":null,"percent":35.4,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":48},{"model":"DeepSeek Coder 33B","score":0.354,"effort":null,"percent":35.4,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":49},{"model":"Qwen2.5-Coder-0.5B","score":0.345,"effort":null,"percent":34.5,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":50,"model_id":"alibaba/qwen2.5-coder-0.5b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-0.5b.toml"},{"model":"MPT-30B","score":0.344,"effort":null,"percent":34.4,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":51},{"model":"Falcon-40B","score":0.338,"effort":null,"percent":33.8,"date":"2023-05-25","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":52},{"model":"PaLM 62B","score":0.33,"effort":null,"percent":33.0,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":53},{"model":"StarCoder 2 7B","score":0.327,"effort":null,"percent":32.7,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":54},{"model":"chatglm2-6b","score":0.324,"effort":null,"percent":32.4,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":55},{"model":"internlm-7b","score":0.312,"effort":null,"percent":31.2,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":56},{"model":"vicuna-13b-v1.1","score":0.2813,"effort":null,"percent":28.1,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":57},{"model":"Baichuan 1-13B","score":0.268,"effort":null,"percent":26.8,"date":"2023-07-11","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":58},{"model":"Baichuan 2-7B","score":0.246,"effort":null,"percent":24.6,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":59},{"model":"Vicuna-13B-v1.3","score":0.226,"effort":null,"percent":22.6,"date":"2023-06-18","organization":"Large Model Systems Organization,University of California (UC) Berkeley","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":60},{"model":"StarCoder 2 3B","score":0.216,"effort":null,"percent":21.6,"date":"2024-02-22","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":61},{"model":"DeepSeek Coder 6.7B","score":0.213,"effort":null,"percent":21.3,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":62},{"model":"Qwen-1.8B","score":0.212,"effort":null,"percent":21.2,"date":"2023-11-30","organization":null,"countries":[],"accessibility":null,"rank":63},{"model":"LLaMA-13B","score":0.2055,"effort":null,"percent":20.5,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":64},{"model":"Gemma 2B","score":0.177,"effort":null,"percent":17.7,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":65},{"model":"Llama 2-7B","score":0.167,"effort":null,"percent":16.7,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":66},{"model":"internlm-chat-20b","score":0.157,"effort":null,"percent":15.7,"date":"2023-09-17","organization":null,"countries":[],"accessibility":null,"rank":67},{"model":"LLaMA-7B","score":0.11,"effort":null,"percent":11.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":68},{"model":"BLOOM-176B","score":0.095,"effort":null,"percent":9.5,"date":"2022-07-06","organization":"Hugging Face,BigScience","countries":["France","United States of America"],"accessibility":"Open weights (restricted use)","rank":69},{"model":"Baichuan1-7B","score":0.092,"effort":null,"percent":9.2,"date":"2023-06-01","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":70},{"model":"MPT-7B","score":0.091,"effort":null,"percent":9.1,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":71},{"model":"GPT-3 175B (davinci)","score":0.09,"effort":null,"percent":9.0,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":72},{"model":"Falcon-7B","score":0.068,"effort":null,"percent":6.8,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":73},{"model":"DeepSeek Coder 1.3B","score":0.044,"effort":null,"percent":4.4,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":74},{"model":"OPT-175B","score":0.04,"effort":null,"percent":4.0,"date":"2022-05-02","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":75},{"model":"OPT-66B","score":0.018,"effort":null,"percent":1.8,"date":"2022-05-03","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":76},{"model":"GPT-3 6.7B","score":0.016,"effort":null,"percent":1.6,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":77},{"model":"GPT-3 XL (babbage)","score":0.007,"effort":null,"percent":0.7,"date":"2020-06-22","organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":78},{"model":"GPT-3 Medium","score":0.006,"effort":null,"percent":0.6,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":79},{"model":"InstructGPT 6B","score":0.006,"effort":null,"percent":0.6,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":80},{"model":"InstructGPT 350M","score":0.004,"effort":null,"percent":0.4,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":81},{"model":"InstructGPT 1.3B","score":0.0,"effort":null,"percent":0.0,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":82}]},{"id":"winogrande","name":"Winogrande","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.5,"released":"2019-07-24","entrants":77,"leader":{"model":"Llama 3.1-405B","score":0.892,"effort":null,"percent":89.2,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":1},"rows":[{"model":"Llama 3.1-405B","score":0.892,"effort":null,"percent":89.2,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":1},{"model":"Claude 3 Opus","score":0.885,"effort":null,"percent":88.5,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"GPT-4 (Mar 2023)","score":0.875,"effort":null,"percent":87.5,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Falcon-180B","score":0.871,"effort":null,"percent":87.1,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":4},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.863,"effort":null,"percent":86.3,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":5},{"model":"DeepSeek-V3","score":0.852,"effort":null,"percent":85.2,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":6,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"PaLM (540B)","score":0.851,"effort":null,"percent":85.1,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":7},{"model":"DeepSeek-Coder-V2 236B","score":0.837,"effort":null,"percent":83.7,"date":"2024-06-17","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":8},{"model":"Llama 3-70B","score":0.835,"effort":null,"percent":83.5,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":9},{"model":"PaLM 2-L","score":0.83,"effort":null,"percent":83.0,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":10},{"model":"Qwen2.5-72B","score":0.823,"effort":null,"percent":82.3,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"davinci-002","score":0.816,"effort":null,"percent":81.6,"date":"2022-03-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"GPT-3.5 Turbo (June 2023)","score":0.816,"effort":null,"percent":81.6,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"phi-3-small 7.4B","score":0.815,"effort":null,"percent":81.5,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":14},{"model":"phi-3-medium 14B","score":0.815,"effort":null,"percent":81.5,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":15},{"model":"Qwen2.5-Coder-32B","score":0.808,"effort":null,"percent":80.8,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16},{"model":"Llama 2-70B","score":0.802,"effort":null,"percent":80.2,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":17},{"model":"PaLM 2-M","score":0.792,"effort":null,"percent":79.2,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":18},{"model":"GLaM","score":0.792,"effort":null,"percent":79.2,"date":"2021-12-13","organization":"Google","countries":["United States of America"],"accessibility":"Unreleased","rank":19},{"model":"Gemma 7B","score":0.79,"effort":null,"percent":79.0,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Megatron-Turing NLG 530B","score":0.789,"effort":null,"percent":78.9,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":21},{"model":"Falcon 2-11B","score":0.783,"effort":null,"percent":78.3,"date":"2024-05-09","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"Nemotron-4 15B","score":0.78,"effort":null,"percent":78.0,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":23},{"model":"PaLM 2-S","score":0.779,"effort":null,"percent":77.9,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":24},{"model":"InstructGPT 175B","score":0.777,"effort":null,"percent":77.7,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"Mixtral 8x7B","score":0.772,"effort":null,"percent":77.2,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":26},{"model":"PaLM 62B","score":0.77,"effort":null,"percent":77.0,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":27},{"model":"LLaMA-65B","score":0.77,"effort":null,"percent":77.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":28},{"model":"Falcon-40B","score":0.769,"effort":null,"percent":76.9,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":29},{"model":"Qwen2.5-Coder-14B","score":0.768,"effort":null,"percent":76.8,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":30},{"model":"Llama 2-34B","score":0.767,"effort":null,"percent":76.7,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":31},{"model":"LLaMA-33B","score":0.76,"effort":null,"percent":76.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":32},{"model":"Llama 3-8B","score":0.757,"effort":null,"percent":75.7,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":33},{"model":"Mistral 7B v0.1","score":0.753,"effort":null,"percent":75.3,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"Claude 3 Sonnet","score":0.751,"effort":null,"percent":75.1,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":35},{"model":"Chinchilla","score":0.749,"effort":null,"percent":74.9,"date":"2022-03-29","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":36},{"model":"Claude 3 Haiku","score":0.742,"effort":null,"percent":74.2,"date":"2024-03-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"Phi-1.5","score":0.734,"effort":null,"percent":73.4,"date":"2023-09-11","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"LLaMA-13B","score":0.73,"effort":null,"percent":73.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":39},{"model":"Yi-9B","score":0.73,"effort":null,"percent":73.0,"date":"2024-03-01","organization":null,"countries":[],"accessibility":null,"rank":40},{"model":"Qwen2.5-Coder (7B)","score":0.729,"effort":null,"percent":72.9,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"DeepSeek-Coder-V2-Lite-Base","score":0.729,"effort":null,"percent":72.9,"date":"2024-06-13","organization":null,"countries":[],"accessibility":null,"rank":42},{"model":"Llama 2-13B","score":0.728,"effort":null,"percent":72.8,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":43},{"model":"Yi 6B","score":0.713,"effort":null,"percent":71.3,"date":"2023-11-02","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":44},{"model":"MPT-30B","score":0.71,"effort":null,"percent":71.0,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":45},{"model":"vicuna-13b-v1.1","score":0.708,"effort":null,"percent":70.8,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":46},{"model":"phi-3-mini 3.8B","score":0.708,"effort":null,"percent":70.8,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":47},{"model":"Gopher (280B)","score":0.702,"effort":null,"percent":70.2,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":48},{"model":"LLaMA-7B","score":0.701,"effort":null,"percent":70.1,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":49},{"model":"Llama 2-7B","score":0.692,"effort":null,"percent":69.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":50},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.688,"effort":null,"percent":68.8,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":51},{"model":"MPT-7B","score":0.686,"effort":null,"percent":68.6,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":52},{"model":"Qwen2.5-Coder-3B","score":0.674,"effort":null,"percent":67.4,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":53},{"model":"Falcon-7B","score":0.672,"effort":null,"percent":67.2,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":54},{"model":"open_llama_7b","score":0.67,"effort":null,"percent":67.0,"date":"2023-06-07","organization":null,"countries":[],"accessibility":null,"rank":55},{"model":"GPT-NeoX-20B","score":0.661,"effort":null,"percent":66.1,"date":"2022-04-07","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":56},{"model":"INTELLECT-1","score":0.6582,"effort":null,"percent":65.8,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":57},{"model":"Gemma 2B","score":0.654,"effort":null,"percent":65.4,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":58},{"model":"XGen-7B","score":0.649,"effort":null,"percent":64.9,"date":"2023-06-27","organization":"Salesforce","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":59},{"model":"opt-13b","score":0.647,"effort":null,"percent":64.7,"date":"2022-05-11","organization":null,"countries":[],"accessibility":null,"rank":60},{"model":"GPT-J-6B","score":0.645,"effort":null,"percent":64.5,"date":"2021-08-05","organization":"EleutherAI,LAION","countries":["Germany","United States of America"],"accessibility":"Open weights (unrestricted)","rank":61},{"model":"StarCoder 2 15B","score":0.643,"effort":null,"percent":64.3,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":62},{"model":"RedPajama-INCITE-7B-Base","score":0.638,"effort":null,"percent":63.8,"date":"2023-05-04","organization":null,"countries":[],"accessibility":null,"rank":63},{"model":"DeepSeek Coder 33B","score":0.62,"effort":null,"percent":62.0,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":64},{"model":"Dolly 2.0-12b","score":0.618,"effort":null,"percent":61.8,"date":"2023-04-11","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":65},{"model":"OPT-1.3B","score":0.61,"effort":null,"percent":61.0,"date":"2022-05-11","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":66},{"model":"Cerebras-GPT-13B","score":0.608,"effort":null,"percent":60.8,"date":"2023-03-20","organization":"Cerebras Systems","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":67},{"model":"Qwen2.5-Coder (1.5B)","score":0.607,"effort":null,"percent":60.7,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":68},{"model":"CodeQwen1.5-7B","score":0.598,"effort":null,"percent":59.8,"date":"2024-04-15","organization":null,"countries":[],"accessibility":null,"rank":69},{"model":"GPT-2 (1.5B)","score":0.583,"effort":null,"percent":58.3,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":70},{"model":"DeepSeek Coder 6.7B","score":0.576,"effort":null,"percent":57.6,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":71},{"model":"StarCoder 2 3B","score":0.571,"effort":null,"percent":57.1,"date":"2024-02-22","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":72},{"model":"StarCoder 2 7B","score":0.571,"effort":null,"percent":57.1,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":73},{"model":"Qwen2.5-Coder-0.5B","score":0.548,"effort":null,"percent":54.8,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":74,"model_id":"alibaba/qwen2.5-coder-0.5b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-0.5b.toml"},{"model":"Phi-2","score":0.547,"effort":null,"percent":54.7,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":75},{"model":"DeepSeek Coder 1.3B","score":0.533,"effort":null,"percent":53.3,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":76},{"model":"stablelm-tuned-alpha-7b","score":0.515,"effort":null,"percent":51.5,"date":"2023-04-19","organization":null,"countries":[],"accessibility":null,"rank":77}]},{"id":"arc-ai2","name":"ARC AI2","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2018-03-14","entrants":76,"leader":{"model":"Llama 3.1-405B","score":0.953,"effort":null,"percent":95.3,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":1},"rows":[{"model":"Llama 3.1-405B","score":0.953,"effort":null,"percent":95.3,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":1},{"model":"DeepSeek-V3","score":0.953,"effort":null,"percent":95.3,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":2,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Qwen2.5-72B","score":0.945,"effort":null,"percent":94.5,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":3},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.922,"effort":null,"percent":92.2,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":4},{"model":"phi-3-medium 14B","score":0.916,"effort":null,"percent":91.6,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":5},{"model":"phi-3-small 7.4B","score":0.907,"effort":null,"percent":90.7,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":6},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.874,"effort":null,"percent":87.4,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Mixtral 8x7B","score":0.873,"effort":null,"percent":87.3,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":8},{"model":"Claude Instant","score":0.863,"effort":null,"percent":86.3,"date":"2023-08-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Stable Beluga 2","score":0.861,"effort":null,"percent":86.1,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":10},{"model":"PaLM (540B)","score":0.852,"effort":null,"percent":85.2,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":11},{"model":"davinci-002","score":0.852,"effort":null,"percent":85.2,"date":"2022-03-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"phi-3-mini 3.8B","score":0.849,"effort":null,"percent":84.9,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":13},{"model":"Qwen-14B","score":0.844,"effort":null,"percent":84.4,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":14},{"model":"Llama 3-8B","score":0.828,"effort":null,"percent":82.8,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":15},{"model":"internlm-20b","score":0.817,"effort":null,"percent":81.7,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":16},{"model":"Mistral 7B v0.1","score":0.786,"effort":null,"percent":78.6,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":17},{"model":"Llama 2-70B","score":0.783,"effort":null,"percent":78.3,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":18},{"model":"Gemma 7B","score":0.783,"effort":null,"percent":78.3,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":19},{"model":"Phi-2","score":0.759,"effort":null,"percent":75.9,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":20},{"model":"Qwen-7B","score":0.753,"effort":null,"percent":75.3,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"Qwen2.5-Coder-32B","score":0.705,"effort":null,"percent":70.5,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"internlm-7b","score":0.695,"effort":null,"percent":69.5,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":23},{"model":"LLaMA-65B","score":0.695,"effort":null,"percent":69.5,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":24},{"model":"PaLM 2-L","score":0.692,"effort":null,"percent":69.2,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":25},{"model":"Falcon-180B","score":0.678,"effort":null,"percent":67.8,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":26},{"model":"LLaMA-33B","score":0.675,"effort":null,"percent":67.5,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":27},{"model":"Qwen2.5-Coder-14B","score":0.66,"effort":null,"percent":66.0,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":28},{"model":"PaLM 2-M","score":0.649,"effort":null,"percent":64.9,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":29},{"model":"DeepSeek-Coder-V2 236B","score":0.643,"effort":null,"percent":64.3,"date":"2024-06-17","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":30},{"model":"Falcon-40B","score":0.6186,"effort":null,"percent":61.9,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":31},{"model":"chatglm2-6b","score":0.61,"effort":null,"percent":61.0,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":32},{"model":"Qwen2.5-Coder (7B)","score":0.609,"effort":null,"percent":60.9,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33},{"model":"Llama 2-13B","score":0.603,"effort":null,"percent":60.3,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":34},{"model":"PaLM 2-S","score":0.596,"effort":null,"percent":59.6,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":35},{"model":"DeepSeek-Coder-V2-Lite-Base","score":0.573,"effort":null,"percent":57.3,"date":"2024-06-13","organization":null,"countries":[],"accessibility":null,"rank":36},{"model":"Yi-9B","score":0.556,"effort":null,"percent":55.6,"date":"2024-03-01","organization":null,"countries":[],"accessibility":null,"rank":37},{"model":"Nemotron-4 15B","score":0.555,"effort":null,"percent":55.5,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":38},{"model":"INTELLECT-1","score":0.5452,"effort":null,"percent":54.5,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":39},{"model":"Llama 2-34B","score":0.545,"effort":null,"percent":54.5,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":40},{"model":"InstructGPT 175B","score":0.532,"effort":null,"percent":53.2,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"Qwen-1.8B","score":0.532,"effort":null,"percent":53.2,"date":"2023-11-30","organization":null,"countries":[],"accessibility":null,"rank":42},{"model":"Qwen2.5-Coder-3B","score":0.529,"effort":null,"percent":52.9,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":43},{"model":"LLaMA-13B","score":0.527,"effort":null,"percent":52.7,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":44},{"model":"PaLM 62B","score":0.525,"effort":null,"percent":52.5,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":45},{"model":"MPT-30B","score":0.506,"effort":null,"percent":50.6,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":46},{"model":"Yi 6B","score":0.503,"effort":null,"percent":50.3,"date":"2023-11-02","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":47},{"model":"Falcon-7B","score":0.4786,"effort":null,"percent":47.9,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":48},{"model":"LLaMA-7B","score":0.476,"effort":null,"percent":47.6,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":49},{"model":"StarCoder 2 15B","score":0.472,"effort":null,"percent":47.2,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":50},{"model":"Llama 2-7B","score":0.459,"effort":null,"percent":45.9,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":51},{"model":"Qwen2.5-Coder (1.5B)","score":0.452,"effort":null,"percent":45.2,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":52},{"model":"Phi-1.5","score":0.444,"effort":null,"percent":44.4,"date":"2023-09-11","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":53},{"model":"vicuna-13b-v1.1","score":0.432,"effort":null,"percent":43.2,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":54},{"model":"MPT-7B","score":0.426,"effort":null,"percent":42.6,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":55},{"model":"DeepSeek Coder 33B","score":0.422,"effort":null,"percent":42.2,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":56},{"model":"Gemma 2B","score":0.421,"effort":null,"percent":42.1,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":57},{"model":"XGen-7B","score":0.412,"effort":null,"percent":41.2,"date":"2023-06-27","organization":"Salesforce","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":58},{"model":"GPT-NeoX-20B","score":0.411,"effort":null,"percent":41.1,"date":"2022-04-07","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":59},{"model":"Dolly 2.0-12b","score":0.396,"effort":null,"percent":39.6,"date":"2023-04-11","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":60},{"model":"RedPajama-INCITE-7B-Base","score":0.391,"effort":null,"percent":39.1,"date":"2023-05-04","organization":null,"countries":[],"accessibility":null,"rank":61},{"model":"StarCoder 2 7B","score":0.387,"effort":null,"percent":38.7,"date":"2024-02-20","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":62},{"model":"open_llama_7b","score":0.387,"effort":null,"percent":38.7,"date":"2023-06-07","organization":null,"countries":[],"accessibility":null,"rank":63},{"model":"Baichuan2-13B","score":0.38,"effort":null,"percent":38.0,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":64},{"model":"DeepSeek Coder 6.7B","score":0.364,"effort":null,"percent":36.4,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":65},{"model":"GPT-J-6B","score":0.363,"effort":null,"percent":36.3,"date":"2021-08-05","organization":"EleutherAI,LAION","countries":["Germany","United States of America"],"accessibility":"Open weights (unrestricted)","rank":66},{"model":"opt-13b","score":0.358,"effort":null,"percent":35.8,"date":"2022-05-11","organization":null,"countries":[],"accessibility":null,"rank":67},{"model":"CodeQwen1.5-7B","score":0.357,"effort":null,"percent":35.7,"date":"2024-04-15","organization":null,"countries":[],"accessibility":null,"rank":68},{"model":"Qwen2.5-Coder-0.5B","score":0.344,"effort":null,"percent":34.4,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":69,"model_id":"alibaba/qwen2.5-coder-0.5b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-0.5b.toml"},{"model":"StarCoder 2 3B","score":0.342,"effort":null,"percent":34.2,"date":"2024-02-22","organization":"Hugging Face,ServiceNow,NVIDIA,BigCode","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":70},{"model":"Baichuan 2-7B","score":0.325,"effort":null,"percent":32.5,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":71},{"model":"Cerebras-GPT-13B","score":0.324,"effort":null,"percent":32.4,"date":"2023-03-20","organization":"Cerebras Systems","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":72},{"model":"stablelm-tuned-alpha-7b","score":0.27,"effort":null,"percent":27.0,"date":"2023-04-19","organization":null,"countries":[],"accessibility":null,"rank":73},{"model":"DeepSeek Coder 1.3B","score":0.254,"effort":null,"percent":25.4,"date":"2023-11-02","organization":"DeepSeek,Peking University","countries":["China"],"accessibility":"Open weights (restricted use)","rank":74},{"model":"GPT-2 (1.5B)","score":0.25,"effort":null,"percent":25.0,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":75},{"model":"OPT-1.3B","score":0.232,"effort":null,"percent":23.2,"date":"2022-05-11","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":76}]},{"id":"frontiermath-tiers-1-3-v2-private","name":"FrontierMath-Tiers-1-3-v2-Private","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-12","entrants":75,"leader":{"model":"GPT-6 Astra","score":0.9368,"effort":"max","percent":93.7,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.9368,"effort":"max","percent":93.7,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5.1","score":0.9018,"effort":"max","percent":90.2,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-5.6 Sol","score":0.8912,"effort":"max","percent":89.1,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.5 Pro","score":0.8772,"effort":"xhigh","percent":87.7,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"Claude Fable 5","score":0.8702,"effort":"max","percent":87.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.6 Terra","score":0.8596,"effort":"max","percent":86.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Opus 5","score":0.8561,"effort":"max","percent":85.6,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.5","score":0.8526,"effort":"xhigh","percent":85.3,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4 Pro","score":0.8246,"effort":"xhigh","percent":82.5,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"GPT-5.6 Luna","score":0.8211,"effort":"max","percent":82.1,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Opus 4.8","score":0.8,"effort":null,"percent":80.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.4","score":0.786,"effort":"xhigh","percent":78.6,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Qwen3.8 Max","score":0.7474,"effort":"xhigh","percent":74.7,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":13,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GPT-5.2 Pro","score":0.74,"effort":null,"percent":74.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"Kimi K3","score":0.7218,"effort":"max","percent":72.2,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":15,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.7 Flash","score":0.7158,"effort":"high","percent":71.6,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Claude Opus 4.7","score":0.7018,"effort":"max","percent":70.2,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GLM-5.3","score":0.6877,"effort":"max","percent":68.8,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":18,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Gemini 3.8 Flash","score":0.6842,"effort":"high","percent":68.4,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"GPT-5.2","score":0.674,"effort":"xhigh","percent":67.4,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Grok 4.6","score":0.6596,"effort":"xhigh","percent":66.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Claude Opus 4.6","score":0.6596,"effort":"max","percent":66.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Qwen3.8 Max (0902)","score":0.6561,"effort":"xhigh","percent":65.6,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":23},{"model":"Claude Sonnet 5","score":0.6561,"effort":"max","percent":65.6,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.6456,"effort":"max","percent":64.6,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":25,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Qwen3.7 Max","score":0.6456,"effort":null,"percent":64.6,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":26,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Gemini 3.5 Flash","score":0.6281,"effort":null,"percent":62.8,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.5965,"effort":null,"percent":59.6,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GLM-5.2","score":0.5921,"effort":null,"percent":59.2,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Gemini 3.6 Flash","score":0.5895,"effort":null,"percent":58.9,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.5754,"effort":"max","percent":57.5,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":31,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Grok 4.5","score":0.5719,"effort":"high","percent":57.2,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Kimi K2.6","score":0.5719,"effort":null,"percent":57.2,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"GLM-5.3-Flash","score":0.5579,"effort":"max","percent":55.8,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":34,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"GPT-5 Pro","score":0.5579,"effort":null,"percent":55.8,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"GPT-5","score":0.5544,"effort":"high","percent":55.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Kimi K2.7 Code","score":0.5404,"effort":null,"percent":54.0,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":37,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5.4 mini","score":0.5123,"effort":"xhigh","percent":51.2,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Gemini 3 Flash Preview","score":0.5123,"effort":null,"percent":51.2,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"GPT-5 mini","score":0.4667,"effort":"high","percent":46.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Inkling Small","score":0.4632,"effort":"xhigh","percent":46.3,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"DeepSeek v4 Pro","score":0.4526,"effort":"max","percent":45.3,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5.4 nano","score":0.4491,"effort":"high","percent":44.9,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Grok 4.20","score":0.4491,"effort":null,"percent":44.9,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"Grok 4.3 Beta","score":0.4281,"effort":null,"percent":42.8,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":45},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.3825,"effort":null,"percent":38.2,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":46},{"model":"GLM-5.1","score":0.3684,"effort":null,"percent":36.8,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"o4-mini","score":0.3614,"effort":"high","percent":36.1,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Qwen3.6 27B","score":0.3509,"effort":null,"percent":35.1,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":49,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"Qwen3.7 Plus","score":0.3439,"effort":null,"percent":34.4,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":50,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.3439,"effort":null,"percent":34.4,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":51},{"model":"o3","score":0.3333,"effort":"high","percent":33.3,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Inkling","score":0.3333,"effort":"xhigh","percent":33.3,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":53},{"model":"Qwen3.5 397B-A17B","score":0.3123,"effort":null,"percent":31.2,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":54,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.2772,"effort":null,"percent":27.7,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"GPT-5.5 Instant","score":0.2632,"effort":null,"percent":26.3,"date":"2026-05-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.2596,"effort":null,"percent":26.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":57,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.2456,"effort":null,"percent":24.6,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":58},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.2386,"effort":null,"percent":23.9,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.2246,"effort":null,"percent":22.5,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":60},{"model":"Qwen3.6 35B-A3B","score":0.2035,"effort":null,"percent":20.4,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":61,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"GPT-5 nano","score":0.2,"effort":"high","percent":20.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":62,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Qwen3.7 Flash","score":0.193,"effort":null,"percent":19.3,"date":"2026-07-27","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":63,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"Qwen3-Max-Instruct","score":0.1895,"effort":null,"percent":18.9,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":64},{"model":"o3-mini","score":0.186,"effort":"high","percent":18.6,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":65,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Qwen3.5 Flash","score":0.1825,"effort":null,"percent":18.2,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":66,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"o1","score":0.1474,"effort":"high","percent":14.7,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":67,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude Opus 4.1","score":0.1263,"effort":null,"percent":12.6,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"GPT-4.1 mini","score":0.0667,"effort":null,"percent":6.7,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":69,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"GPT-4.1","score":0.0596,"effort":null,"percent":6.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":70,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"GPT-5.4 nano (no thinking)","score":0.0456,"effort":null,"percent":4.6,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":71},{"model":"GPT-4o mini","score":0.007,"effort":null,"percent":0.7,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":72,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-4 Turbo (Apr 2024)","score":0.007,"effort":null,"percent":0.7,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73},{"model":"GPT-4o (Aug 2024)","score":0.0035,"effort":null,"percent":0.4,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":74},{"model":"GPT-3.5 Turbo (Jan 2024)","score":0.0,"effort":null,"percent":0.0,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":75}]},{"id":"hellaswag","name":"HellaSwag","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2019-05-19","entrants":75,"leader":{"model":"GPT-4 (Mar 2023)","score":0.953,"effort":null,"percent":95.3,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1},"rows":[{"model":"GPT-4 (Mar 2023)","score":0.953,"effort":null,"percent":95.3,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1},{"model":"Llama 3.1-405B","score":0.892,"effort":null,"percent":89.2,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":2},{"model":"Falcon-180B","score":0.89,"effort":null,"percent":89.0,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":3},{"model":"DeepSeek-V3","score":0.889,"effort":null,"percent":88.9,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":4,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.871,"effort":null,"percent":87.1,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":5},{"model":"PaLM 2-L","score":0.868,"effort":null,"percent":86.8,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":6},{"model":"Mixtral 8x7B","score":0.867,"effort":null,"percent":86.7,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"davinci-003","score":0.855,"effort":null,"percent":85.5,"date":"2022-11-28","organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":8},{"model":"Llama 2-70B","score":0.853,"effort":null,"percent":85.3,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":9},{"model":"Falcon-40B","score":0.8528,"effort":null,"percent":85.3,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":10},{"model":"Qwen2.5-72B","score":0.848,"effort":null,"percent":84.8,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"LLaMA-65B","score":0.842,"effort":null,"percent":84.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":12},{"model":"Stable Beluga 2","score":0.841,"effort":null,"percent":84.1,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":13},{"model":"PaLM 2-M","score":0.84,"effort":null,"percent":84.0,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":14},{"model":"PaLM (540B)","score":0.838,"effort":null,"percent":83.8,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":15},{"model":"Qwen2.5-Coder-32B","score":0.83,"effort":null,"percent":83.0,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16},{"model":"Falcon 2-11B","score":0.8291,"effort":null,"percent":82.9,"date":"2024-05-09","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":17},{"model":"LLaMA-33B","score":0.828,"effort":null,"percent":82.8,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":18},{"model":"Megatron-Turing NLG 530B","score":0.824,"effort":null,"percent":82.4,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":19},{"model":"Nemotron-4 15B","score":0.824,"effort":null,"percent":82.4,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":20},{"model":"phi-3-medium 14B","score":0.824,"effort":null,"percent":82.4,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":21},{"model":"Gemma 7B","score":0.822,"effort":null,"percent":82.2,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"PaLM 2-S","score":0.82,"effort":null,"percent":82.0,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":23},{"model":"davinci-002","score":0.815,"effort":null,"percent":81.5,"date":"2022-03-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24},{"model":"Mistral 7B v0.1","score":0.81,"effort":null,"percent":81.0,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":25},{"model":"Llama 2-13B","score":0.807,"effort":null,"percent":80.7,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":26},{"model":"Qwen2.5-Coder-14B","score":0.802,"effort":null,"percent":80.2,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":27},{"model":"InstructGPT 175B","score":0.793,"effort":null,"percent":79.3,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Gopher (280B)","score":0.792,"effort":null,"percent":79.2,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":29},{"model":"LLaMA-13B","score":0.792,"effort":null,"percent":79.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":30},{"model":"OPT-175B","score":0.791,"effort":null,"percent":79.1,"date":"2022-05-02","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":31},{"model":"Falcon-7B","score":0.781,"effort":null,"percent":78.1,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"internlm-20b","score":0.781,"effort":null,"percent":78.1,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":33},{"model":"GPT-3 175B (davinci)","score":0.775,"effort":null,"percent":77.5,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34},{"model":"GLaM","score":0.772,"effort":null,"percent":77.2,"date":"2021-12-13","organization":"Google","countries":["United States of America"],"accessibility":"Unreleased","rank":35},{"model":"Llama 2-7B","score":0.772,"effort":null,"percent":77.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":36},{"model":"phi-3-small 7.4B","score":0.77,"effort":null,"percent":77.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":37},{"model":"Qwen2.5-Coder (7B)","score":0.768,"effort":null,"percent":76.8,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"phi-3-mini 3.8B","score":0.767,"effort":null,"percent":76.7,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":39},{"model":"MPT-7B","score":0.764,"effort":null,"percent":76.4,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":40},{"model":"Yi-9B","score":0.764,"effort":null,"percent":76.4,"date":"2024-03-01","organization":null,"countries":[],"accessibility":null,"rank":41},{"model":"LLaMA-7B","score":0.762,"effort":null,"percent":76.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":42},{"model":"OPT-66B","score":0.745,"effort":null,"percent":74.5,"date":"2022-05-03","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":43},{"model":"BLOOM-176B","score":0.744,"effort":null,"percent":74.4,"date":"2022-07-06","organization":"Hugging Face,BigScience","countries":["France","United States of America"],"accessibility":"Open weights (restricted use)","rank":44},{"model":"Yi 6B","score":0.744,"effort":null,"percent":74.4,"date":"2023-11-02","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":45},{"model":"XGen-7B","score":0.742,"effort":null,"percent":74.2,"date":"2023-06-27","organization":"Salesforce","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":46},{"model":"open_llama_7b","score":0.718,"effort":null,"percent":71.8,"date":"2023-06-07","organization":null,"countries":[],"accessibility":null,"rank":47},{"model":"INTELLECT-1","score":0.7142,"effort":null,"percent":71.4,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":48},{"model":"Gemma 2B","score":0.714,"effort":null,"percent":71.4,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":49},{"model":"Qwen2.5-Coder-3B","score":0.709,"effort":null,"percent":70.9,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":50},{"model":"Baichuan2-13B","score":0.708,"effort":null,"percent":70.8,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":51},{"model":"Dolly 2.0-12b","score":0.708,"effort":null,"percent":70.8,"date":"2023-04-11","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":52},{"model":"internlm-7b","score":0.706,"effort":null,"percent":70.6,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":53},{"model":"GPT-NeoX-20B","score":0.705,"effort":null,"percent":70.5,"date":"2022-04-07","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":54},{"model":"RedPajama-INCITE-7B-Base","score":0.703,"effort":null,"percent":70.3,"date":"2023-05-04","organization":null,"countries":[],"accessibility":null,"rank":55},{"model":"opt-13b","score":0.699,"effort":null,"percent":69.9,"date":"2022-05-11","organization":null,"countries":[],"accessibility":null,"rank":56},{"model":"GPT-3 6.7B","score":0.682,"effort":null,"percent":68.2,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":57},{"model":"Baichuan 2-7B","score":0.68,"effort":null,"percent":68.0,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":58},{"model":"InstructGPT 6B","score":0.676,"effort":null,"percent":67.6,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":59},{"model":"GPT-J-6B","score":0.662,"effort":null,"percent":66.2,"date":"2021-08-05","organization":"EleutherAI,LAION","countries":["Germany","United States of America"],"accessibility":"Open weights (unrestricted)","rank":60},{"model":"Qwen2.5-Coder (1.5B)","score":0.618,"effort":null,"percent":61.8,"date":"2024-09-18","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":61},{"model":"Cerebras-GPT-13B","score":0.594,"effort":null,"percent":59.4,"date":"2023-03-20","organization":"Cerebras Systems","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":62},{"model":"vicuna-13b-v1.1","score":0.578,"effort":null,"percent":57.8,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":63},{"model":"chatglm2-6b","score":0.57,"effort":null,"percent":57.0,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":64},{"model":"InstructGPT 1.3B","score":0.561,"effort":null,"percent":56.1,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":65},{"model":"GPT-3 XL (babbage)","score":0.555,"effort":null,"percent":55.5,"date":"2020-06-22","organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":66},{"model":"Phi-2","score":0.536,"effort":null,"percent":53.6,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":67},{"model":"Qwen2.5-Coder-0.5B","score":0.484,"effort":null,"percent":48.4,"date":"2024-09-18","organization":null,"countries":[],"accessibility":null,"rank":68,"model_id":"alibaba/qwen2.5-coder-0.5b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-0.5b.toml"},{"model":"Phi-1.5","score":0.476,"effort":null,"percent":47.6,"date":"2023-09-11","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":69},{"model":"GPT-3 Medium","score":0.435,"effort":null,"percent":43.5,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":70},{"model":"InstructGPT 350M","score":0.429,"effort":null,"percent":42.9,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":null,"rank":71},{"model":"GPT-Neo-2.7B","score":0.427,"effort":null,"percent":42.7,"date":"2021-03-30","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":72},{"model":"OPT-1.3B","score":0.415,"effort":null,"percent":41.5,"date":"2022-05-11","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":73},{"model":"stablelm-tuned-alpha-7b","score":0.407,"effort":null,"percent":40.7,"date":"2023-04-19","organization":null,"countries":[],"accessibility":null,"rank":74},{"model":"GPT-2 (1.5B)","score":0.4,"effort":null,"percent":40.0,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":75}]},{"id":"mystery-game-puzzles","name":"Mystery Game Puzzles","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0922,"released":"2026-07-30","entrants":74,"leader":{"model":"GPT-6 Astra","score":0.84,"effort":"max","percent":84.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.84,"effort":"max","percent":84.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Opus 5","score":0.59,"effort":"max","percent":59.0,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Fable 5.1","score":0.58,"effort":"max","percent":58.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-5.6 Sol","score":0.58,"effort":"max","percent":58.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.5","score":0.56,"effort":"xhigh","percent":56.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Fable 5","score":0.52,"effort":"max","percent":52.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Gemini 3.8 Flash","score":0.47,"effort":"high","percent":47.0,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.43,"effort":"max","percent":43.0,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":8,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Qwen3.8 Max","score":0.38,"effort":"xhigh","percent":38.0,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":9,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GPT-5.4","score":0.37,"effort":"xhigh","percent":37.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.7 Flash","score":0.37,"effort":"high","percent":37.0,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Claude Opus 4.8","score":0.36,"effort":null,"percent":36.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.6 Terra","score":0.35,"effort":"max","percent":35.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Sonnet 5","score":0.35,"effort":"max","percent":35.0,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Grok 4.6","score":0.34,"effort":"xhigh","percent":34.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.34,"effort":null,"percent":34.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.34,"effort":"max","percent":34.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"GLM-5.3","score":0.33,"effort":"max","percent":33.0,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":18,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Gemini 3.5 Flash","score":0.32,"effort":null,"percent":32.0,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Qwen3.7 Max","score":0.32,"effort":null,"percent":32.0,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":20,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Gemini 3.6 Flash","score":0.3,"effort":null,"percent":30.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"o3","score":0.29,"effort":"high","percent":29.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude Opus 4.7","score":0.28,"effort":"max","percent":28.0,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Kimi K3","score":0.26,"effort":"max","percent":26.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":24,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Qwen3.7-Max","score":0.26,"effort":null,"percent":26.0,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":25,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Gemini 3 Flash Preview","score":0.26,"effort":null,"percent":26.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Claude Opus 4.6","score":0.25,"effort":"max","percent":25.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.2","score":0.23,"effort":"high","percent":23.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"GPT-5","score":0.23,"effort":"high","percent":23.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Qwen3.6 35B-A3B","score":0.22,"effort":null,"percent":22.0,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"Claude Opus 4.5 (48k thinking)","score":0.22,"effort":null,"percent":22.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"GPT-5.6 Luna","score":0.21,"effort":"max","percent":21.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Opus 4.1","score":0.21,"effort":null,"percent":21.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Qwen3.5 Flash","score":0.2,"effort":null,"percent":20.0,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":34,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"Nemotron 3 Ultra","score":0.2,"effort":null,"percent":20.0,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":35},{"model":"Qwen 3.6 Max (Preview)","score":0.19,"effort":null,"percent":19.0,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":36},{"model":"GPT-5.1","score":0.19,"effort":"low","percent":19.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"GLM-5.2","score":0.19,"effort":null,"percent":19.0,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Gemini 3.5 Flash-Lite","score":0.19,"effort":null,"percent":19.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Kimi K2.6","score":0.18,"effort":null,"percent":18.0,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":40,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.18,"effort":null,"percent":18.0,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":41},{"model":"Qwen3.5 397B-A17B","score":0.18,"effort":null,"percent":18.0,"date":"2026-02-13","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42,"model_id":"alibaba/qwen3.5-397b-a17b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-397b-a17b.toml"},{"model":"GPT-5.5 (no thinking)","score":0.18,"effort":null,"percent":18.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43},{"model":"Qwen3.7 Plus","score":0.17,"effort":null,"percent":17.0,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":44,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"Qwen3.5-122B-A10B","score":0.17,"effort":null,"percent":17.0,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":45,"model_id":"alibaba/qwen3.5-122b-a10b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-122b-a10b.toml"},{"model":"Qwen3.5 Plus","score":0.17,"effort":null,"percent":17.0,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":46,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"DeepSeek v4 Pro","score":0.17,"effort":"unknown thinking","percent":17.0,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Claude Sonnet 4.5 (48k thinking)","score":0.17,"effort":null,"percent":17.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"Claude Sonnet 4.6","score":0.16,"effort":"low","percent":16.0,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Sonnet 5 (no thinking)","score":0.16,"effort":null,"percent":16.0,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":50},{"model":"Qwen3.7 Flash","score":0.15,"effort":null,"percent":15.0,"date":"2026-07-27","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":51,"model_id":"alibaba/qwen3.7-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-flash.toml"},{"model":"GPT-5.1 (no thinking)","score":0.15,"effort":null,"percent":15.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"Claude Opus 4.6 (no thinking)","score":0.15,"effort":null,"percent":15.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"Claude Sonnet 4.6 (no thinking)","score":0.14,"effort":null,"percent":14.0,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":54},{"model":"Claude Opus 4.7 (no thinking)","score":0.13,"effort":null,"percent":13.0,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":55},{"model":"GPT-4 (Jun 2023)","score":0.12,"effort":null,"percent":12.0,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.12,"effort":null,"percent":12.0,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":57},{"model":"GPT-4o mini","score":0.12,"effort":null,"percent":12.0,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":58,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-5.4 mini","score":0.11,"effort":"none","percent":11.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":59,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GPT-5 mini","score":0.1,"effort":"minimal","percent":10.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":60,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GPT-5 nano","score":0.09,"effort":"medium","percent":9.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":61,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.09,"effort":null,"percent":9.0,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":62},{"model":"GPT-5.4 nano (no thinking)","score":0.09,"effort":null,"percent":9.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":63},{"model":"GLM-5.3-Flash","score":0.08,"effort":"max","percent":8.0,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":64,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"MiniMax-M3","score":0.08,"effort":null,"percent":8.0,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":65},{"model":"o3-mini","score":0.07,"effort":"high","percent":7.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":66,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Qwen3.6 27B","score":0.07,"effort":null,"percent":7.0,"date":"2026-04-22","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":67,"model_id":"alibaba/qwen3.6-27b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-27b.toml"},{"model":"GPT-4.1 mini","score":0.07,"effort":null,"percent":7.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"GPT-5.4 nano","score":0.06,"effort":"medium","percent":6.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":69,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Inkling Small","score":0.06,"effort":"xhigh","percent":6.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":70},{"model":"o4-mini","score":0.05,"effort":"high","percent":5.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":71,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Qwen3-Max-Instruct","score":0.05,"effort":null,"percent":5.0,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":72},{"model":"GPT-3.5 Turbo (Jan 2024)","score":0.03,"effort":null,"percent":3.0,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":73},{"model":"gpt-oss-120b","score":0.02,"effort":null,"percent":2.0,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":74,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"}]},{"id":"simpleqa-verified","name":"SimpleQA Verified","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-09-09","entrants":72,"leader":{"model":"GPT-6 Astra","score":0.756,"effort":"max","percent":75.6,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.756,"effort":"max","percent":75.6,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.735,"effort":null,"percent":73.5,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Fable 5.1","score":0.708,"effort":"max","percent":70.8,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Fable 5","score":0.707,"effort":"xhigh","percent":70.7,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Gemini 3.8 Flash","score":0.697,"effort":"high","percent":69.7,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"GPT-5.6 Sol","score":0.697,"effort":"max","percent":69.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Gemini 3.7 Flash","score":0.692,"effort":"high","percent":69.2,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Gemini 3 Flash Preview","score":0.668,"effort":null,"percent":66.8,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Gemini 3.5 Flash","score":0.662,"effort":null,"percent":66.2,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Gemini 3.6 Flash","score":0.662,"effort":null,"percent":66.2,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GPT-5.5","score":0.63,"effort":"xhigh","percent":63.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Muse Spark 1.2","score":0.603,"effort":"xhigh","percent":60.3,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"Claude Opus 5","score":0.599,"effort":"max","percent":59.9,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Muse Spark 1.1","score":0.5779,"effort":null,"percent":57.8,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"Qwen3.7 Max","score":0.5577,"effort":null,"percent":55.8,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":15,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Claude Opus 4.8","score":0.53,"effort":null,"percent":53.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.5291,"effort":"max","percent":52.9,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Qwen 3.6 Max (Preview)","score":0.52,"effort":null,"percent":52.0,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":18},{"model":"Claude Opus 4.7","score":0.517,"effort":"xhigh","percent":51.7,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Kimi K3","score":0.506,"effort":"max","percent":50.6,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":20,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5","score":0.501,"effort":"high","percent":50.1,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o3","score":0.494,"effort":"high","percent":49.4,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Grok 4.6","score":0.493,"effort":"high","percent":49.3,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Qwen3-Max-Instruct","score":0.4875,"effort":null,"percent":48.7,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":24},{"model":"Grok 4.5","score":0.483,"effort":"high","percent":48.3,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"GPT-5.1","score":0.48,"effort":"high","percent":48.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Qwen3.8 Max (0902)","score":0.473,"effort":"xhigh","percent":47.3,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":27},{"model":"Claude Opus 4.6","score":0.47,"effort":"max","percent":47.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"DeepSeek v4 Pro","score":0.4699,"effort":"max","percent":47.0,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5.4 Pro","score":0.463,"effort":"xhigh","percent":46.3,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Qwen3.8 Max","score":0.458,"effort":"xhigh","percent":45.8,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":31,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.457,"effort":null,"percent":45.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"GPT-5.4","score":0.451,"effort":"xhigh","percent":45.1,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.4414,"effort":null,"percent":44.1,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":34},{"model":"GPT-5.6 Terra","score":0.432,"effort":"max","percent":43.2,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"o1","score":0.411,"effort":"high","percent":41.1,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"GLM-5.3","score":0.41,"effort":"max","percent":41.0,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":37,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"GPT-5.6 Luna","score":0.41,"effort":"max","percent":41.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.4044,"effort":null,"percent":40.4,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":39},{"model":"Inkling","score":0.403,"effort":"xhigh","percent":40.3,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":40},{"model":"GPT-5.2","score":0.371,"effort":"xhigh","percent":37.1,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Kimi K2.7 Code","score":0.365,"effort":null,"percent":36.5,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Claude Sonnet 4.6","score":0.355,"effort":"high","percent":35.5,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Kimi K2.6","score":0.349,"effort":null,"percent":34.9,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Kimi K2.5","score":0.343,"effort":null,"percent":34.3,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":45},{"model":"GLM-5.2","score":0.342,"effort":null,"percent":34.2,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":46,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"GLM-5.1","score":0.34,"effort":null,"percent":34.0,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Claude Sonnet 5","score":0.337,"effort":"max","percent":33.7,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.3363,"effort":"max","percent":33.6,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":49,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Grok 4.3 Beta","score":0.332,"effort":null,"percent":33.2,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":50},{"model":"GLM-4.7","score":0.322,"effort":null,"percent":32.2,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":51,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"GPT-4.1","score":0.311,"effort":null,"percent":31.1,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Claude Sonnet 4.5 (59k thinking)","score":0.307,"effort":null,"percent":30.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"Grok 4.20","score":0.302,"effort":null,"percent":30.2,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":54},{"model":"GPT-5.4 mini","score":0.294,"effort":"high","percent":29.4,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GPT-4o (Aug 2024)","score":0.26,"effort":null,"percent":26.0,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"Qwen3.5 Plus","score":0.2535,"effort":null,"percent":25.4,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":57,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.237,"effort":null,"percent":23.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":58},{"model":"GPT-5 mini","score":0.216,"effort":"high","percent":21.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":59,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Qwen3.5 Flash","score":0.2032,"effort":null,"percent":20.3,"date":"2026-02-25","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":60,"model_id":"alibaba/qwen3.5-flash","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-flash.toml"},{"model":"o4-mini","score":0.196,"effort":"low","percent":19.6,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":61,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Inkling Small","score":0.191,"effort":"xhigh","percent":19.1,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":62},{"model":"Qwen 3.6 Flash (2026-04-16)","score":0.1592,"effort":null,"percent":15.9,"date":"2026-04-27","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":63},{"model":"o3-mini","score":0.153,"effort":"high","percent":15.3,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":64,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Claude Haiku 4.5","score":0.132,"effort":null,"percent":13.2,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":65,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"GPT-4.1 mini","score":0.127,"effort":null,"percent":12.7,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":66,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Claude 3 Opus","score":0.126,"effort":null,"percent":12.6,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":67},{"model":"GPT-5.4 nano","score":0.117,"effort":"high","percent":11.7,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":68,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"GPT-5 nano","score":0.117,"effort":"high","percent":11.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":69,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Gemma 4 31B IT","score":0.104,"effort":null,"percent":10.4,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":70,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"GPT-4o mini","score":0.083,"effort":null,"percent":8.3,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":71,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"GPT-4.1 nano","score":0.06,"effort":null,"percent":6.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":72,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"}]},{"id":"proofbench","name":"ProofBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-01-30","entrants":64,"leader":{"model":"Claude Fable 5.1","score":1.0,"effort":"max","percent":100.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},"rows":[{"model":"Claude Fable 5.1","score":1.0,"effort":"max","percent":100.0,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Opus 5","score":0.99,"effort":"max","percent":99.0,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-6 Astra","score":0.99,"effort":"unknown thinking","percent":99.0,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5","score":0.95,"effort":"max","percent":95.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Kimi K3","score":0.87,"effort":"unknown","percent":87.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":5,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.6 Sol","score":0.83,"effort":"max","percent":83.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Sonnet 5","score":0.77,"effort":"max","percent":77.0,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"GPT-5.6 Terra","score":0.74,"effort":"xhigh","percent":74.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Claude Opus 4.8","score":0.69,"effort":null,"percent":69.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.6 Luna","score":0.6,"effort":"max","percent":60.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Gemini 3.7 Flash","score":0.58,"effort":"unknown","percent":58.0,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Qwen3.8 Max","score":0.58,"effort":"unknown","percent":58.0,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":12,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GPT-5.4","score":0.56,"effort":"xhigh","percent":56.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.56,"effort":"unknown","percent":56.0,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Claude Opus 4.7","score":0.54,"effort":"max","percent":54.0,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Grok 4.6","score":0.51,"effort":"unknown","percent":51.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.5","score":0.5,"effort":"xhigh","percent":50.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Opus 4.6","score":0.5,"effort":"max","percent":50.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.5,"effort":"unknown","percent":50.0,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"GLM-5.3","score":0.49,"effort":"max","percent":49.0,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":20,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Gemini 3.8 Flash","score":0.48,"effort":"unknown","percent":48.0,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Claude Sonnet 4.6","score":0.45,"effort":"max","percent":45.0,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Muse Spark 1.2","score":0.43,"effort":null,"percent":43.0,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"Muse Spark 1.1","score":0.39,"effort":null,"percent":39.0,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"Claude Opus 4.5","score":0.36,"effort":"unknown thinking","percent":36.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Gemini 3.6 Flash","score":0.36,"effort":null,"percent":36.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GLM-5.2","score":0.35,"effort":null,"percent":35.0,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Grok 4.5","score":0.31,"effort":"high","percent":31.0,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Gemini 3.5 Flash","score":0.31,"effort":null,"percent":31.0,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Qwen3.7 Max","score":0.26,"effort":null,"percent":26.0,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":30,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.26,"effort":null,"percent":26.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GLM-5.1","score":0.2222,"effort":null,"percent":22.2,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":32,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"MiMo-V2.5-Pro","score":0.22,"effort":null,"percent":22.0,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":33},{"model":"GPT-5.4 mini","score":0.21,"effort":"xhigh","percent":21.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GLM-5.3-Flash","score":0.21,"effort":"max","percent":21.0,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":35,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Gemini 3 Pro Preview","score":0.2,"effort":null,"percent":20.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Claude Sonnet 4.5","score":0.19,"effort":"unknown thinking","percent":19.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"MiniMax-M3","score":0.18,"effort":null,"percent":18.0,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":38},{"model":"GPT-5","score":0.18,"effort":"high","percent":18.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Muse Spark","score":0.17,"effort":null,"percent":17.0,"date":"2026-04-08","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Kimi K2.6","score":0.16,"effort":null,"percent":16.0,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"MiMo-V2.5","score":0.16,"effort":null,"percent":16.0,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":42},{"model":"DeepSeek v4 Pro","score":0.16,"effort":"max","percent":16.0,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Qwen3.8 27B","score":0.16,"effort":null,"percent":16.0,"date":"2026-08-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44,"model_id":"alibaba/qwen3.8-27b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-27b.toml"},{"model":"Gemini 3 Flash Preview","score":0.15,"effort":null,"percent":15.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"GPT-5.2","score":0.15,"effort":"xhigh","percent":15.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Grok 4.20","score":0.14,"effort":null,"percent":14.0,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"Gemini 3.5 Flash-Lite","score":0.13,"effort":null,"percent":13.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"GPT-5 nano","score":0.12,"effort":"high","percent":12.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Grok 4.3 Beta","score":0.11,"effort":null,"percent":11.0,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":50},{"model":"GPT-5 mini","score":0.09,"effort":"high","percent":9.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":51,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GPT-5.1-Codex-Max","score":0.09,"effort":null,"percent":9.0,"date":"2025-11-19","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/gpt-5.1-codex-max","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex-max.toml"},{"model":"Mistral Medium 3.5","score":0.09,"effort":null,"percent":9.0,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":53,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"DeepSeek-V3.2 (Thinking; Fireworks)","score":0.08,"effort":null,"percent":8.0,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":54},{"model":"GLM-4.7","score":0.06,"effort":null,"percent":6.0,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":55,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"Inkling Small","score":0.06,"effort":null,"percent":6.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":56},{"model":"GPT-5.4 nano","score":0.05,"effort":"high","percent":5.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":57,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Grok 4.1 Fast","score":0.04,"effort":null,"percent":4.0,"date":"2025-11-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":58,"model_id":"xai/grok-4.1-fast","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.1-fast.toml"},{"model":"MiniMax-M2.5","score":0.04,"effort":null,"percent":4.0,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":59},{"model":"MiniMax-M2.7","score":0.03,"effort":null,"percent":3.0,"date":"2026-03-18","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":60},{"model":"Nemotron 3 Ultra","score":0.02,"effort":null,"percent":2.0,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":61},{"model":"Laguna XS.2","score":0.0,"effort":null,"percent":0.0,"date":"2026-04-28","organization":"Poolside","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":62},{"model":"Laguna M.1","score":0.0,"effort":null,"percent":0.0,"date":"2026-04-28","organization":"Poolside","countries":["United States of America"],"accessibility":null,"rank":63},{"model":"Inkling","score":0.0,"effort":null,"percent":0.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":64}]},{"id":"piqa","name":"PIQA","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.5,"released":"2019-11-26","entrants":59,"leader":{"model":"GPT-4o mini","score":0.887,"effort":null,"percent":88.7,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},"rows":[{"model":"GPT-4o mini","score":0.887,"effort":null,"percent":88.7,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Phi-3.5-MoE","score":0.886,"effort":null,"percent":88.6,"date":"2024-08-17","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":2},{"model":"Gemini 1.5 Flash (Sep 2024)","score":0.875,"effort":null,"percent":87.5,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Llama 3.1-405B","score":0.859,"effort":null,"percent":85.9,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":4},{"model":"Falcon-180B","score":0.849,"effort":null,"percent":84.9,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":5},{"model":"DeepSeek-V3","score":0.847,"effort":null,"percent":84.7,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":6,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Inflection-1","score":0.842,"effort":null,"percent":84.2,"date":"2023-06-22","organization":"Inflection AI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":7},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.839,"effort":null,"percent":83.9,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":8},{"model":"Gemma 2 9B","score":0.837,"effort":null,"percent":83.7,"date":null,"organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":9},{"model":"Mixtral 8x7B","score":0.836,"effort":null,"percent":83.6,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":10},{"model":"Mistral NeMo","score":0.835,"effort":null,"percent":83.5,"date":"2024-07-18","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":11,"model_id":"mistral/mistral-nemo","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-nemo.toml"},{"model":"Stable Beluga 2","score":0.833,"effort":null,"percent":83.3,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":12},{"model":"Megatron-Turing NLG 530B","score":0.8319,"effort":null,"percent":83.2,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":13},{"model":"Mistral 7B v0.1","score":0.83,"effort":null,"percent":83.0,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":14},{"model":"Falcon-40B","score":0.83,"effort":null,"percent":83.0,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":15},{"model":"LLaMA-65B","score":0.828,"effort":null,"percent":82.8,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":16},{"model":"Llama 2-70B","score":0.828,"effort":null,"percent":82.8,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":17},{"model":"Qwen2.5-72B","score":0.826,"effort":null,"percent":82.6,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18},{"model":"Nemotron-4 15B","score":0.824,"effort":null,"percent":82.4,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":19},{"model":"LLaMA-33B","score":0.823,"effort":null,"percent":82.3,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":20},{"model":"InstructGPT 175B","score":0.823,"effort":null,"percent":82.3,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"PaLM (540B)","score":0.823,"effort":null,"percent":82.3,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":22},{"model":"Mistral 7B v0.2","score":0.822,"effort":null,"percent":82.2,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":23},{"model":"MPT-30B","score":0.819,"effort":null,"percent":81.9,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":24},{"model":"Llama 2-34B","score":0.819,"effort":null,"percent":81.9,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":25},{"model":"Gopher (280B)","score":0.818,"effort":null,"percent":81.8,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":26},{"model":"Chinchilla","score":0.818,"effort":null,"percent":81.8,"date":"2022-03-29","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":27},{"model":"Gemma 7B","score":0.812,"effort":null,"percent":81.2,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":28},{"model":"Llama 3.1-8B","score":0.812,"effort":null,"percent":81.2,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":29},{"model":"Phi-3.5-mini","score":0.81,"effort":null,"percent":81.0,"date":"2024-08-16","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":30},{"model":"Llama 2-13B","score":0.808,"effort":null,"percent":80.8,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":31},{"model":"MPT-7B","score":0.806,"effort":null,"percent":80.6,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"PaLM 62B","score":0.805,"effort":null,"percent":80.5,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":33},{"model":"Falcon-7B","score":0.803,"effort":null,"percent":80.3,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"internlm-20b","score":0.803,"effort":null,"percent":80.3,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":35},{"model":"LLaMA-13B","score":0.801,"effort":null,"percent":80.1,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":36},{"model":"Qwen-14B","score":0.799,"effort":null,"percent":79.9,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":37},{"model":"LLaMA-7B","score":0.798,"effort":null,"percent":79.8,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":38},{"model":"Llama 2-7B","score":0.788,"effort":null,"percent":78.8,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":39},{"model":"Baichuan2-13B","score":0.781,"effort":null,"percent":78.1,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":40},{"model":"internlm-7b","score":0.779,"effort":null,"percent":77.9,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":41},{"model":"Qwen-7B","score":0.779,"effort":null,"percent":77.9,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":42},{"model":"vicuna-13b-v1.1","score":0.774,"effort":null,"percent":77.4,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":43},{"model":"Gemma 2B","score":0.773,"effort":null,"percent":77.3,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":44},{"model":"RedPajama-INCITE-7B-Base","score":0.769,"effort":null,"percent":76.9,"date":"2023-05-04","organization":null,"countries":[],"accessibility":null,"rank":45},{"model":"GPT-NeoX-20B","score":0.767,"effort":null,"percent":76.7,"date":"2022-04-07","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":46},{"model":"Baichuan1-7B","score":0.762,"effort":null,"percent":76.2,"date":"2023-06-01","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":47},{"model":"open_llama_7b","score":0.76,"effort":null,"percent":76.0,"date":"2023-06-07","organization":null,"countries":[],"accessibility":null,"rank":48},{"model":"opt-13b","score":0.757,"effort":null,"percent":75.7,"date":"2022-05-11","organization":null,"countries":[],"accessibility":null,"rank":49},{"model":"XGen-7B","score":0.755,"effort":null,"percent":75.5,"date":"2023-06-27","organization":"Salesforce","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":50},{"model":"GPT-J-6B","score":0.754,"effort":null,"percent":75.4,"date":"2021-08-05","organization":"EleutherAI,LAION","countries":["Germany","United States of America"],"accessibility":"Open weights (unrestricted)","rank":51},{"model":"Dolly 2.0-12b","score":0.754,"effort":null,"percent":75.4,"date":"2023-04-11","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":52},{"model":"Cerebras-GPT-13B","score":0.735,"effort":null,"percent":73.5,"date":"2023-03-20","organization":"Cerebras Systems","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":53},{"model":"Qwen-1.8B","score":0.733,"effort":null,"percent":73.3,"date":"2023-11-30","organization":null,"countries":[],"accessibility":null,"rank":54},{"model":"GPT-Neo-2.7B","score":0.729,"effort":null,"percent":72.9,"date":"2021-03-30","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":55},{"model":"GPT-2 (1.5B)","score":0.705,"effort":null,"percent":70.5,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":56},{"model":"chatglm2-6b","score":0.696,"effort":null,"percent":69.6,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":57},{"model":"OPT-1.3B","score":0.69,"effort":null,"percent":69.0,"date":"2022-05-11","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":58},{"model":"stablelm-tuned-alpha-7b","score":0.658,"effort":null,"percent":65.8,"date":"2023-04-19","organization":null,"countries":[],"accessibility":null,"rank":59}]},{"id":"fiction-livebench","name":"Fiction.LiveBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-02-21","entrants":58,"leader":{"model":"GPT-5","score":0.972,"effort":"medium","percent":97.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},"rows":[{"model":"GPT-5","score":0.972,"effort":"medium","percent":97.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o3-pro","score":0.972,"effort":null,"percent":97.2,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"Grok 4","score":0.944,"effort":null,"percent":94.4,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Grok 4 Fast","score":0.944,"effort":null,"percent":94.4,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.917,"effort":null,"percent":91.7,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"o3","score":0.889,"effort":"medium","percent":88.9,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Kimi K2.5","score":0.861,"effort":null,"percent":86.1,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"QwQ-32B","score":0.833,"effort":null,"percent":83.3,"date":"2025-03-05","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":8,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"model":"Claude 3.7 Sonnet (8k thinking)","score":0.833,"effort":null,"percent":83.3,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"DeepSeek-V3.2-Exp","score":0.833,"effort":null,"percent":83.3,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":10},{"model":"o1","score":0.833,"effort":"medium","percent":83.3,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":0.778,"effort":null,"percent":77.8,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"o4-mini","score":0.778,"effort":"medium","percent":77.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"DeepSeek-R1 (May 2025)","score":0.75,"effort":null,"percent":75.0,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":0.75,"effort":null,"percent":75.0,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":15},{"model":"Qwen3-32B","score":0.742,"effort":null,"percent":74.2,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16},{"model":"DeepSeek-R1","score":0.694,"effort":null,"percent":69.4,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"GPT-5 mini","score":0.694,"effort":"medium","percent":69.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"MiniMax-M1-80k","score":0.694,"effort":null,"percent":69.4,"date":"2025-06-13","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Qwen3-235B-A22B","score":0.677,"effort":null,"percent":67.7,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":20},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.667,"effort":null,"percent":66.7,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"GPT-4o (Jan 2025)","score":0.667,"effort":null,"percent":66.7,"date":"2025-01-29","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22},{"model":"Kimi K2 Instruct (0905)","score":0.667,"effort":null,"percent":66.7,"date":"2025-09-05","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":23},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.667,"effort":null,"percent":66.7,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":24},{"model":"Grok-3 mini","score":0.667,"effort":null,"percent":66.7,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"Qwen2.5-Max","score":0.667,"effort":null,"percent":66.7,"date":"2025-01-25","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":26},{"model":"Qwen3-Max-Instruct","score":0.667,"effort":null,"percent":66.7,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":27},{"model":"GPT-4.1","score":0.639,"effort":null,"percent":63.9,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.639,"effort":null,"percent":63.9,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29},{"model":"Qwen3-14B","score":0.625,"effort":null,"percent":62.5,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30},{"model":"Qwen3-8B","score":0.621,"effort":null,"percent":62.1,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":31},{"model":"Claude Opus 4","score":0.611,"effort":null,"percent":61.1,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.611,"effort":null,"percent":61.1,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":33},{"model":"Kimi K2 Instruct","score":0.611,"effort":null,"percent":61.1,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":34},{"model":"chutes/GLM-4.5-FP8","score":0.583,"effort":null,"percent":58.3,"date":"2025-07-27","organization":null,"countries":[],"accessibility":null,"rank":35},{"model":"Grok 3","score":0.583,"effort":null,"percent":58.3,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"deepinfra/Qwen3-Next-80B-A3B-Instruct","score":0.556,"effort":null,"percent":55.6,"date":"2025-09-11","organization":null,"countries":[],"accessibility":null,"rank":37},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":0.529,"effort":null,"percent":52.9,"date":"2025-09-01","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"Gemini 2.0 Flash Thinking Exp","score":0.528,"effort":null,"percent":52.8,"date":"2025-01-21","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"DeepSeek-V3.1","score":0.528,"effort":null,"percent":52.8,"date":"2025-08-21","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":40,"model_id":"deepseek/deepseek-v3.1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.1.toml"},{"model":"Claude 3.7 Sonnet","score":0.5,"effort":null,"percent":50.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"o3-mini","score":0.5,"effort":"medium","percent":50.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":42,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"DeepSeek-V3 (Mar 2025)","score":0.5,"effort":null,"percent":50.0,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":43},{"model":"gemini-2.5-flash-lite-preview-06-17 (Default thinking length)","score":0.472,"effort":null,"percent":47.2,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":0.472,"effort":null,"percent":47.2,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":45},{"model":"Claude Sonnet 4","score":0.469,"effort":null,"percent":46.9,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Llama 4 Maverick","score":0.462,"effort":null,"percent":46.2,"date":"2025-04-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":47},{"model":"gpt-oss-120b","score":0.444,"effort":null,"percent":44.4,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":48,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"GPT-4.1 mini","score":0.444,"effort":null,"percent":44.4,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":49,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"GPT-5 nano","score":0.444,"effort":"medium","percent":44.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Gemini 2.0 Pro Exp (Feb 2025)","score":0.417,"effort":null,"percent":41.7,"date":"2025-02-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":51},{"model":"chutes/Qwen3-Next-80B-A3B-Instruct","score":0.417,"effort":null,"percent":41.7,"date":"2025-09-11","organization":null,"countries":[],"accessibility":null,"rank":52},{"model":"Qwen3-30B-A3B","score":0.406,"effort":null,"percent":40.6,"date":"2025-04-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":53,"model_id":"alibaba/qwen3-30b-a3b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-30b-a3b.toml"},{"model":"Llama 4 Scout","score":0.36,"effort":null,"percent":36.0,"date":"2025-04-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":54},{"model":"Gemma 3 27B","score":0.333,"effort":null,"percent":33.3,"date":"2025-03-11","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":55},{"model":"Llama 3.3 70B","score":0.333,"effort":null,"percent":33.3,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":56},{"model":"nvidia-nemotron-nano-9b-v2","score":0.25,"effort":null,"percent":25.0,"date":"2025-08-18","organization":null,"countries":[],"accessibility":null,"rank":57},{"model":"GPT-4.1 nano","score":0.25,"effort":null,"percent":25.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":58,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"}]},{"id":"frontiermath-tier-4-v2-private","name":"FrontierMath-Tier-4-v2-Private","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-12","entrants":57,"leader":{"model":"GPT-6 Astra","score":0.976,"effort":"high","percent":97.6,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.976,"effort":"high","percent":97.6,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5","score":0.902,"effort":"max","percent":90.2,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Fable 5.1","score":0.878,"effort":"max","percent":87.8,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-5.6 Sol","score":0.8293,"effort":"max","percent":82.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":0.8049,"effort":null,"percent":80.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"GPT-5.5 Pro","score":0.7805,"effort":"xhigh","percent":78.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.5-pro","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-pro.toml"},{"model":"AI co-mathematician","score":0.756,"effort":null,"percent":75.6,"date":"2026-05-08","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Unreleased","rank":7},{"model":"Claude Opus 5","score":0.7317,"effort":"max","percent":73.2,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.5","score":0.725,"effort":"xhigh","percent":72.5,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.6 Terra","score":0.7073,"effort":"max","percent":70.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.6 Luna","score":0.6098,"effort":"max","percent":61.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"GPT-5.4 Pro","score":0.5854,"effort":"xhigh","percent":58.5,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Claude Opus 4.8","score":0.561,"effort":null,"percent":56.1,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.4","score":0.49,"effort":"xhigh","percent":49.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Qwen3.8 Max","score":0.4634,"effort":"xhigh","percent":46.3,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":15,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"GPT-5.2 Pro","score":0.46,"effort":null,"percent":46.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"openai/gpt-5.2-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-pro.toml"},{"model":"Kimi K3","score":0.3902,"effort":"max","percent":39.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":17,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.7 Flash","score":0.3659,"effort":"high","percent":36.6,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Qwen3.8 Max (0902)","score":0.3415,"effort":"xhigh","percent":34.1,"date":"2026-09-01","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":19},{"model":"Qwen3.7 Max","score":0.3415,"effort":null,"percent":34.1,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":20,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Grok 4.6","score":0.3171,"effort":"xhigh","percent":31.7,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Claude Opus 4.7","score":0.3171,"effort":"max","percent":31.7,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.2","score":0.317,"effort":"xhigh","percent":31.7,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"GLM-5.3","score":0.2927,"effort":"max","percent":29.3,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":24,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Claude Sonnet 5","score":0.2927,"effort":"max","percent":29.3,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"GLM-5.2","score":0.2927,"effort":null,"percent":29.3,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"DeepSeek V4 Pro 0813","score":0.2683,"effort":"max","percent":26.8,"date":"2026-08-13","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27,"model_id":"deepseek/deepseek-v4-pro-0813","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro-0813.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.2683,"effort":null,"percent":26.8,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Opus 4.6","score":0.2683,"effort":"max","percent":26.8,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Gemini 3.5 Flash","score":0.2683,"effort":null,"percent":26.8,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Kimi K2.6","score":0.2564,"effort":null,"percent":25.6,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":31,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.2439,"effort":"max","percent":24.4,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":32,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"Grok 4.5","score":0.2439,"effort":"high","percent":24.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Gemini 3.8 Flash","score":0.2195,"effort":"high","percent":22.0,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Gemini 3.6 Flash","score":0.2195,"effort":null,"percent":22.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GPT-5","score":0.2195,"effort":"high","percent":22.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5 Pro","score":0.1951,"effort":null,"percent":19.5,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"GLM-5.3-Flash","score":0.1707,"effort":"max","percent":17.1,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Inkling Small","score":0.1707,"effort":"xhigh","percent":17.1,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":39},{"model":"Grok 4.20","score":0.1707,"effort":null,"percent":17.1,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Gemini 3 Flash Preview","score":0.1707,"effort":null,"percent":17.1,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":41,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Grok 4.3 Beta","score":0.1463,"effort":null,"percent":14.6,"date":"2026-04-17","organization":"xAI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":42},{"model":"Kimi K2.7 Code","score":0.122,"effort":null,"percent":12.2,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5 mini","score":0.122,"effort":"high","percent":12.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":44,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GPT-5.4 nano","score":0.122,"effort":"high","percent":12.2,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"GPT-5.4 mini","score":0.0976,"effort":"xhigh","percent":9.8,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":46,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Inkling","score":0.0488,"effort":"xhigh","percent":4.9,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":47},{"model":"o4-mini","score":0.0488,"effort":"high","percent":4.9,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":48,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Opus 4.5 (32k thinking)","score":0.0488,"effort":null,"percent":4.9,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":49},{"model":"GPT-5.5 Instant","score":0.0244,"effort":null,"percent":2.4,"date":"2026-05-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":50,"model_id":"openai/gpt-5.5-instant","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5-instant.toml"},{"model":"DeepSeek v4 Pro","score":0.0244,"effort":"max","percent":2.4,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":51,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"GPT-5 nano","score":0.0244,"effort":"high","percent":2.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":52,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"Claude Opus 4.1","score":0.0244,"effort":null,"percent":2.4,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":53,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.0244,"effort":null,"percent":2.4,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":54},{"model":"Gemini 3.5 Flash-Lite","score":0.0,"effort":null,"percent":0.0,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"google/gemini-3.5-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash-lite.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.0,"effort":null,"percent":0.0,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":56},{"model":"o3-mini","score":0.0,"effort":"high","percent":0.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":57,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"}]},{"id":"aider-polyglot","name":"Aider polyglot","in_index":true,"scale":0.01,"ceiling":1.0,"random_baseline":0.0,"released":"2024-12-21","entrants":55,"leader":{"model":"GPT-5","score":88.0,"effort":"high","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},"rows":[{"model":"GPT-5","score":88.0,"effort":"high","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o3-pro","score":84.9,"effort":null,"percent":null,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":83.1,"effort":null,"percent":null,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"o3","score":81.3,"effort":"high","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Grok 4","score":79.6,"effort":null,"percent":null,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"DeepSeek-V3.2-Exp","score":74.2,"effort":null,"percent":null,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6},{"model":"DeepSeek-V3.2","score":74.2,"effort":"thinking","percent":null,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":7,"model_id":"deepseek/deepseek-v3.2","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.2.toml"},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":72.9,"effort":null,"percent":null,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":72.9,"effort":null,"percent":null,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Claude Opus 4","score":72.0,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"o4-mini","score":72.0,"effort":"high","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"DeepSeek-R1 (May 2025)","score":71.4,"effort":null,"percent":null,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":12},{"model":"Claude 3.7 Sonnet (32k thinking)","score":64.9,"effort":null,"percent":null,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"o1","score":61.7,"effort":"high","percent":null,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude Sonnet 4","score":61.3,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.7 Sonnet","score":60.4,"effort":null,"percent":null,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"o3-mini","score":60.4,"effort":"high","percent":null,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Qwen3-235B-A22B","score":59.6,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18},{"model":"Qwen3-235B-A22B-Instruct (Jul 2025)","score":59.6,"effort":null,"percent":null,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Kimi K2 Instruct","score":59.1,"effort":null,"percent":null,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Kimi K2 0905 (Novita)","score":59.1,"effort":null,"percent":null,"date":"2025-09-05","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"DeepSeek-R1","score":56.9,"effort":null,"percent":null,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":55.1,"effort":null,"percent":null,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"DeepSeek-V3 (Mar 2025)","score":55.1,"effort":null,"percent":null,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":24},{"model":"Grok 3","score":53.3,"effort":null,"percent":null,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"GPT-4.1","score":52.4,"effort":null,"percent":null,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":51.6,"effort":null,"percent":null,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27},{"model":"Grok-3 mini","score":49.3,"effort":null,"percent":null,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"DeepSeek-V3","score":48.4,"effort":null,"percent":null,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":29,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":47.1,"effort":null,"percent":null,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"GPT-4o (Mar 2025)","score":45.3,"effort":null,"percent":null,"date":"2025-03-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"GPT-4.5 Preview (Feb 2025)","score":44.9,"effort":null,"percent":null,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"gpt-oss-120b","score":41.8,"effort":null,"percent":null,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":33,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Qwen3-32B","score":40.0,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"Gemini 2.0 Flash (Dec 2024)","score":38.2,"effort":null,"percent":null,"date":"2024-12-06","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":35},{"model":"Gemini 2.0 Pro Exp (Feb 2025)","score":35.6,"effort":null,"percent":null,"date":"2025-02-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":36},{"model":"o1-mini","score":32.9,"effort":"unknown thinking","percent":null,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"GPT-4.1 mini","score":32.4,"effort":null,"percent":null,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":38,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Claude 3.5 Haiku (Oct 2024)","score":28.0,"effort":null,"percent":null,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"GPT-4o (Jan 2025)","score":27.1,"effort":null,"percent":null,"date":"2025-01-29","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"GPT-4o (Aug 2024)","score":23.1,"effort":null,"percent":null,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"Qwen2.5-Max","score":21.8,"effort":null,"percent":null,"date":"2025-01-25","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":42},{"model":"QwQ-32B","score":20.9,"effort":null,"percent":null,"date":"2025-03-05","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"model":"Gemini 2.0 Flash Thinking Exp","score":18.2,"effort":null,"percent":null,"date":"2025-01-21","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"GPT-4o (Nov 2024)","score":18.2,"effort":null,"percent":null,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":45},{"model":"DeepSeek-V2.5 (Sep 2024)","score":17.8,"effort":null,"percent":null,"date":"2024-09-06","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":46},{"model":"Qwen2.5-Coder-32B-Instruct","score":16.4,"effort":null,"percent":null,"date":"2024-11-21","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":47,"model_id":"alibaba/qwen2.5-coder-32b-instruct","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen2.5-coder-32b-instruct.toml"},{"model":"Llama 4 Maverick","score":15.6,"effort":null,"percent":null,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":48},{"model":"Yi-Lightning","score":12.9,"effort":null,"percent":null,"date":"2024-12-02","organization":"01.AI","countries":["China"],"accessibility":"API access","rank":49},{"model":"Cohere Command A","score":12.0,"effort":null,"percent":null,"date":"2025-03-13","organization":"Cohere","countries":["Canada"],"accessibility":"Open weights (non-commercial)","rank":50},{"model":"Codestral","score":11.1,"effort":null,"percent":null,"date":"2025-01-13","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":51},{"model":"openhands-lm-32b-v0.1","score":10.2,"effort":null,"percent":null,"date":"2024-03-26","organization":null,"countries":[],"accessibility":null,"rank":52},{"model":"GPT-4.1 nano","score":8.9,"effort":null,"percent":null,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":53,"model_id":"openai/gpt-4.1-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-nano.toml"},{"model":"Gemma 3 27B","score":4.9,"effort":null,"percent":null,"date":"2025-03-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":54},{"model":"GPT-4o mini","score":3.6,"effort":null,"percent":null,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"}]},{"id":"apex-agents","name":"APEX-Agents","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-01-21","entrants":55,"leader":{"model":"Claude Fable 5.1","score":0.474,"effort":"unknown","percent":47.4,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},"rows":[{"model":"Claude Fable 5.1","score":0.474,"effort":"unknown","percent":47.4,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"GPT-6 Astra","score":0.467,"effort":"unknown thinking","percent":46.7,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5","score":0.45,"effort":null,"percent":45.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":0.435,"effort":"max","percent":43.5,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Opus 4.8","score":0.425,"effort":null,"percent":42.5,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Muse Spark 1.1","score":0.419,"effort":null,"percent":41.9,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"Grok 4.6","score":0.412,"effort":"unknown","percent":41.2,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.6 Sol (pro, max)","score":0.4,"effort":null,"percent":40.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"GPT-5.6 Sol","score":0.399,"effort":"max","percent":39.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Kimi K3","score":0.393,"effort":"unknown","percent":39.3,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":10,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.5","score":0.385,"effort":"xhigh","percent":38.5,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4","score":0.36,"effort":"xhigh","percent":36.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GLM-5.2","score":0.356,"effort":null,"percent":35.6,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":13,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"GPT-5.2","score":0.344,"effort":"xhigh","percent":34.4,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Grok 4.5","score":0.342,"effort":"unknown thinking","percent":34.2,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Claude Opus 4.7","score":0.339,"effort":"max","percent":33.9,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.335,"effort":null,"percent":33.5,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Sonnet 5","score":0.325,"effort":"unknown thinking","percent":32.5,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Claude Opus 4.6","score":0.324,"effort":"unknown thinking","percent":32.4,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.3 Codex","score":0.318,"effort":null,"percent":31.8,"date":"2026-02-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"Gemini 3 Pro Preview","score":0.315,"effort":null,"percent":31.5,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.2 Codex","score":0.276,"effort":null,"percent":27.6,"date":"2025-12-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/gpt-5.2-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-codex.toml"},{"model":"Kimi K2.7 Code","score":0.276,"effort":null,"percent":27.6,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5.4 mini","score":0.246,"effort":"unknown thinking","percent":24.6,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Gemini 3 Flash Preview","score":0.24,"effort":null,"percent":24.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Claude Sonnet 4.6","score":0.237,"effort":"high","percent":23.7,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"GPT-5.1 Codex","score":0.207,"effort":null,"percent":20.7,"date":"2025-11-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"openai/gpt-5.1-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex.toml"},{"model":"Claude Opus 4.5","score":0.207,"effort":"unknown thinking","percent":20.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"GPT-5-codex","score":0.201,"effort":"high","percent":20.1,"date":"2025-09-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"openai/gpt-5-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-codex.toml"},{"model":"Kimi K2.6","score":0.189,"effort":null,"percent":18.9,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"GPT-5","score":0.183,"effort":"unknown thinking","percent":18.3,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5.1","score":0.175,"effort":"unknown thinking","percent":17.5,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"o3","score":0.172,"effort":"high","percent":17.2,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"GLM-5","score":0.172,"effort":null,"percent":17.2,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":34,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5.4 nano","score":0.169,"effort":"unknown thinking","percent":16.9,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5.4-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-nano.toml"},{"model":"Grok 4","score":0.152,"effort":null,"percent":15.2,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"Kimi K2.5","score":0.144,"effort":null,"percent":14.4,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":37},{"model":"Qwen3.5 Plus","score":0.136,"effort":null,"percent":13.6,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":38,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.13,"effort":null,"percent":13.0,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"Grok 4.1","score":0.128,"effort":null,"percent":12.8,"date":"2025-11-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Nemotron 3 Ultra","score":0.115,"effort":null,"percent":11.5,"date":"2026-06-04","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"Claude Sonnet 4","score":0.093,"effort":"unknown thinking","percent":9.3,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":42,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude Haiku 4.5","score":0.089,"effort":"unknown thinking","percent":8.9,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":43,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"GLM-4.7 (Novita)","score":0.087,"effort":null,"percent":8.7,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44},{"model":"GLM-4.7","score":0.087,"effort":null,"percent":8.7,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":45,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.07,"effort":null,"percent":7.0,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":46},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.066,"effort":null,"percent":6.6,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"MiniMax-M2.5","score":0.062,"effort":null,"percent":6.2,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":48},{"model":"gpt-oss-120b","score":0.047,"effort":null,"percent":4.7,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":49,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Kimi K2 Thinking","score":0.041,"effort":null,"percent":4.1,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":50,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"GLM-4.6","score":0.04,"effort":null,"percent":4.0,"date":"2025-09-30","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":51,"model_id":"zhipuai/glm-4.6","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.6.toml"},{"model":"Grok 3","score":0.021,"effort":null,"percent":2.1,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":52},{"model":"Gemini 2.5 Flash (Jun 2025)","score":0.018,"effort":null,"percent":1.8,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":53},{"model":"o1","score":0.011,"effort":"high","percent":1.1,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":54,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"GPT-4o (Nov 2024)","score":0.011,"effort":null,"percent":1.1,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":55}]},{"id":"terminal-bench","name":"Terminal Bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-05-19","entrants":50,"leader":{"model":"GPT-5.5","score":0.847,"effort":"unknown thinking","percent":84.7,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},"rows":[{"model":"GPT-5.5","score":0.847,"effort":"unknown thinking","percent":84.7,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4","score":0.818,"effort":"unknown thinking","percent":81.8,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.802,"effort":null,"percent":80.2,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Opus 4.7","score":0.802,"effort":"unknown","percent":80.2,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Claude Opus 4.6","score":0.798,"effort":"unknown thinking","percent":79.8,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.3 Codex","score":0.784,"effort":null,"percent":78.4,"date":"2026-02-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.699,"effort":null,"percent":69.9,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Gemini 3 Pro Preview","score":0.694,"effort":null,"percent":69.4,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.2 Codex","score":0.665,"effort":null,"percent":66.5,"date":"2025-12-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.2-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2-codex.toml"},{"model":"GPT-5.2","score":0.649,"effort":"medium","percent":64.9,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3 Flash Preview","score":0.643,"effort":null,"percent":64.3,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.631,"effort":null,"percent":63.1,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Claude Opus 4.5","score":0.631,"effort":"unknown thinking","percent":63.1,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"GPT-5.1 Codex Mini","score":0.616,"effort":null,"percent":61.6,"date":"2025-11-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.1-codex-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex-mini.toml"},{"model":"GPT-5.1-Codex-Max","score":0.604,"effort":null,"percent":60.4,"date":"2025-11-19","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5.1-codex-max","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex-max.toml"},{"model":"Claude Opus 4.5 (128k thinking)","score":0.591,"effort":null,"percent":59.1,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"GPT-5.1 Codex","score":0.578,"effort":null,"percent":57.8,"date":"2025-11-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5.1-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex.toml"},{"model":"Grok 4.20","score":0.573,"effort":null,"percent":57.3,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"Claude Sonnet 4.6","score":0.534,"effort":"unknown thinking","percent":53.4,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Sonnet 4.6 (no thinking)","score":0.534,"effort":null,"percent":53.4,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"GLM-5","score":0.524,"effort":null,"percent":52.4,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5","score":0.496,"effort":"unknown thinking","percent":49.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5.1","score":0.476,"effort":"unknown thinking","percent":47.6,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Claude Sonnet 4.5","score":0.465,"effort":"unknown thinking","percent":46.5,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"MiniMax-M2.7","score":0.451,"effort":null,"percent":45.1,"date":"2026-03-18","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":25},{"model":"GPT-5-codex","score":0.443,"effort":null,"percent":44.3,"date":"2025-09-15","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/gpt-5-codex","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-codex.toml"},{"model":"Kimi K2.5","score":0.432,"effort":null,"percent":43.2,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.428,"effort":null,"percent":42.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"MiniMax-M2.5","score":0.427,"effort":null,"percent":42.7,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.396,"effort":null,"percent":39.6,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":30},{"model":"Claude Opus 4.1","score":0.38,"effort":null,"percent":38.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"MiniMax-M2.1","score":0.366,"effort":null,"percent":36.6,"date":"2025-12-23","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":32},{"model":"Kimi K2 Thinking","score":0.357,"effort":null,"percent":35.7,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":33,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"Claude Haiku 4.5","score":0.355,"effort":"unknown thinking","percent":35.5,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":34,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"GPT-5 mini","score":0.348,"effort":"unknown thinking","percent":34.8,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":35,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GLM-4.7","score":0.334,"effort":null,"percent":33.4,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":36,"model_id":"zhipuai/glm-4.7","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.7.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.326,"effort":null,"percent":32.6,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"MiniMax-M2","score":0.3,"effort":null,"percent":30.0,"date":"2025-10-27","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"Kimi K2 Instruct","score":0.278,"effort":null,"percent":27.8,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":39},{"model":"Grok 4","score":0.272,"effort":null,"percent":27.2,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Qwen3-Coder-480B-A35B","score":0.272,"effort":null,"percent":27.2,"date":"2025-07-31","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"grok-code-fast-1","score":0.258,"effort":null,"percent":25.8,"date":"2025-08-28","organization":null,"countries":[],"accessibility":null,"rank":42},{"model":"GLM-4.6","score":0.245,"effort":null,"percent":24.5,"date":"2025-09-30","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":43,"model_id":"zhipuai/glm-4.6","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.6.toml"},{"model":"Qwen3.6 35B-A3B","score":0.23,"effort":null,"percent":23.0,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":44,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"GPT-5 nano","score":0.218,"effort":"unknown thinking","percent":21.8,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":45,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"gpt-oss-120b","score":0.187,"effort":null,"percent":18.7,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":46,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Gemini 2.5 Flash (Sep 2025)","score":0.171,"effort":null,"percent":17.1,"date":"2025-09-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":47},{"model":"Gemini 2.5 Flash (Jun 2025)","score":0.171,"effort":null,"percent":17.1,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":48},{"model":"Qwen3.5 9B","score":0.092,"effort":null,"percent":9.2,"date":"2026-02-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":49,"model_id":"alibaba/qwen3.5-9b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-9b.toml"},{"model":"gpt-oss-20b","score":0.034,"effort":null,"percent":3.4,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":50,"model_id":"openai/gpt-oss-20b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-20b.toml"}]},{"id":"bbh","name":"BBH","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2022-10-17","entrants":45,"leader":{"model":"Gemini 1.5 Pro (May 2024)","score":0.892,"effort":null,"percent":89.2,"date":"2024-05-14","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1},"rows":[{"model":"Gemini 1.5 Pro (May 2024)","score":0.892,"effort":null,"percent":89.2,"date":"2024-05-14","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1},{"model":"DeepSeek-V3","score":0.875,"effort":null,"percent":87.5,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":2,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Gemini 1.5 Pro (Feb 2024)","score":0.84,"effort":null,"percent":84.0,"date":"2024-02-15","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Llama 3.1-405B","score":0.829,"effort":null,"percent":82.9,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":4},{"model":"phi-3-medium 14B","score":0.814,"effort":null,"percent":81.4,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":5},{"model":"Qwen2.5-72B","score":0.798,"effort":null,"percent":79.8,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6},{"model":"phi-3-small 7.4B","score":0.791,"effort":null,"percent":79.1,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.788,"effort":null,"percent":78.8,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":8},{"model":"GPT-4 (Jun 2023)","score":0.7512,"effort":null,"percent":75.1,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Yi-34B","score":0.717,"effort":null,"percent":71.7,"date":"2023-11-22","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":10},{"model":"phi-3-mini 3.8B","score":0.717,"effort":null,"percent":71.7,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"Stable Beluga 2","score":0.693,"effort":null,"percent":69.3,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":12},{"model":"Llama 2-70B","score":0.649,"effort":null,"percent":64.9,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":13},{"model":"GPT-3.5 Turbo (June 2023)","score":0.6159,"effort":null,"percent":61.6,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"Phi-2","score":0.594,"effort":null,"percent":59.4,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":15},{"model":"Nemotron-4 15B","score":0.587,"effort":null,"percent":58.7,"date":"2024-02-26","organization":"NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":16},{"model":"LLaMA-65B","score":0.584,"effort":null,"percent":58.4,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":17},{"model":"Llama 2-13B","score":0.582,"effort":null,"percent":58.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":18},{"model":"Mistral 7B v0.1","score":0.561,"effort":null,"percent":56.1,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Gemma 7B","score":0.551,"effort":null,"percent":55.1,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Qwen-14B","score":0.55,"effort":null,"percent":55.0,"date":"2023-09-24","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"internlm-20b","score":0.525,"effort":null,"percent":52.5,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":22},{"model":"LLaMA-33B","score":0.5,"effort":null,"percent":50.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":23},{"model":"Baichuan2-13B","score":0.49,"effort":null,"percent":49.0,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":24},{"model":"Yi 6B","score":0.472,"effort":null,"percent":47.2,"date":"2023-11-22","organization":"01.AI","countries":["China"],"accessibility":"Open weights (restricted use)","rank":25},{"model":"Baichuan2-13B-Chat","score":0.472,"effort":null,"percent":47.2,"date":"2023-09-06","organization":null,"countries":[],"accessibility":null,"rank":26},{"model":"Qwen-7B","score":0.45,"effort":null,"percent":45.0,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":27},{"model":"Llama 2-34B","score":0.441,"effort":null,"percent":44.1,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":28},{"model":"vicuna-13b-v1.1","score":0.4304,"effort":null,"percent":43.0,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":29},{"model":"Baichuan 1-13B","score":0.4301,"effort":null,"percent":43.0,"date":"2023-07-11","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":30},{"model":"internlm-chat-20b","score":0.424,"effort":null,"percent":42.4,"date":"2023-09-17","organization":null,"countries":[],"accessibility":null,"rank":31},{"model":"Baichuan 2-7B","score":0.416,"effort":null,"percent":41.6,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":32},{"model":"Llama 2-7B","score":0.3916,"effort":null,"percent":39.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":33},{"model":"MPT-30B","score":0.38,"effort":null,"percent":38.0,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"LLaMA-13B","score":0.379,"effort":null,"percent":37.9,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":35},{"model":"Falcon-40B","score":0.371,"effort":null,"percent":37.1,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":36},{"model":"internlm-7b","score":0.37,"effort":null,"percent":37.0,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":37},{"model":"MPT-7B","score":0.356,"effort":null,"percent":35.6,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"Gemma 2B","score":0.352,"effort":null,"percent":35.2,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":39},{"model":"INTELLECT-1","score":0.3485,"effort":null,"percent":34.8,"date":"2024-11-29","organization":"Prime Intellect,Hugging Face,Arcee AI","countries":["United States of America"],"accessibility":null,"rank":40},{"model":"chatglm2-6b","score":0.337,"effort":null,"percent":33.7,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":41},{"model":"LLaMA-7B","score":0.335,"effort":null,"percent":33.5,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":42},{"model":"Baichuan1-7B","score":0.3248,"effort":null,"percent":32.5,"date":"2023-06-01","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":43},{"model":"Falcon-7B","score":0.2877,"effort":null,"percent":28.8,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":44},{"model":"Qwen-1.8B","score":0.282,"effort":null,"percent":28.2,"date":"2023-11-30","organization":null,"countries":[],"accessibility":null,"rank":45}]},{"id":"lech-mazur-writing","name":"Lech Mazur Writing","in_index":true,"scale":0.1,"ceiling":1.0,"random_baseline":0.0,"released":"2025-01-31","entrants":45,"leader":{"model":"GPT-5","score":8.6,"effort":"medium","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},"rows":[{"model":"GPT-5","score":8.6,"effort":"medium","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Kimi K2 Instruct","score":8.56,"effort":null,"percent":null,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":2},{"model":"Claude Opus 4.1","score":8.47,"effort":null,"percent":null,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"o3-pro","score":8.44,"effort":null,"percent":null,"date":"2025-06-10","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/o3-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-pro.toml"},{"model":"o3","score":8.39,"effort":"medium","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":8.38,"effort":null,"percent":null,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"Claude Opus 4","score":8.36,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"GPT-5 mini","score":8.31,"effort":"medium","percent":null,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Qwen3-235B-A22B","score":8.3,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"DeepSeek-R1","score":8.3,"effort":null,"percent":null,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":10,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Qwen3-235B-A22B-Thinking (Jul 2025)","score":8.24,"effort":null,"percent":null,"date":"2025-07-25","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"DeepSeek-R1 (May 2025)","score":8.19,"effort":null,"percent":null,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":12},{"model":"GPT-4o (Nov 2024)","score":8.18,"effort":null,"percent":null,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"Claude Sonnet 4","score":8.14,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.7 Sonnet (16k thinking)","score":8.11,"effort":null,"percent":null,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":8.09,"effort":null,"percent":null,"date":"2025-05-06","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":8.05,"effort":null,"percent":null,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":8.03,"effort":null,"percent":null,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"QwQ-32B","score":8.02,"effort":null,"percent":null,"date":"2025-03-05","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19,"model_id":"alibaba/qwq-32b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwq-32b.toml"},{"model":"Gemma 3 27B","score":7.99,"effort":null,"percent":null,"date":"2025-03-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Claude 3.7 Sonnet","score":7.94,"effort":null,"percent":null,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"Mistral Medium 3","score":7.73,"effort":null,"percent":null,"date":"2025-05-07","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":22,"model_id":"mistral/mistral-medium-2505","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2505.toml"},{"model":"gpt-oss-120b","score":7.71,"effort":null,"percent":null,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"DeepSeek-V3 (Mar 2025)","score":7.7,"effort":null,"percent":null,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":24},{"model":"Grok 4","score":7.69,"effort":null,"percent":null,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"Gemini 2.5 Flash (Apr 2025)","score":7.65,"effort":null,"percent":null,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":26},{"model":"Grok 3","score":7.64,"effort":null,"percent":null,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":27},{"model":"GPT-4.5 Preview (Feb 2025)","score":7.56,"effort":null,"percent":null,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Qwen3-30B-A3B","score":7.53,"effort":null,"percent":null,"date":"2025-04-29","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"alibaba/qwen3-30b-a3b","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3-30b-a3b.toml"},{"model":"o4-mini","score":7.5,"effort":"medium","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Gemini 2.0 Flash Thinking Exp","score":7.38,"effort":null,"percent":null,"date":"2025-01-21","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"Claude 3.5 Haiku (Oct 2024)","score":7.35,"effort":null,"percent":null,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"Grok-3 mini","score":7.35,"effort":null,"percent":null,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":33},{"model":"GLM-4.5","score":7.34,"effort":null,"percent":null,"date":"2025-08-03","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":34,"model_id":"zhipuai/glm-4.5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.5.toml"},{"model":"Qwen2.5-Max","score":7.29,"effort":null,"percent":null,"date":"2025-01-28","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":35},{"model":"Gemini 2.0 Flash (Dec 2024)","score":7.15,"effort":null,"percent":null,"date":"2024-12-11","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"o1","score":7.02,"effort":"medium","percent":null,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Mistral Large 2 (Jul 2024)","score":6.9,"effort":null,"percent":null,"date":"2024-07-24","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (non-commercial)","rank":38},{"model":"GPT-4o mini","score":6.72,"effort":null,"percent":null,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":39,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"o1-mini","score":6.49,"effort":"medium","percent":null,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Grok-2 (Dec 2024)","score":6.36,"effort":null,"percent":null,"date":"2024-12-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"Phi-4","score":6.26,"effort":null,"percent":null,"date":"2024-12-12","organization":"Microsoft Research","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":42},{"model":"Llama 4 Maverick","score":6.2,"effort":null,"percent":null,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":43},{"model":"o3-mini","score":6.17,"effort":"high","percent":null,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":44,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Amazon Nova Pro","score":6.05,"effort":null,"percent":null,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":45}]},{"id":"metr-time-horizons","name":"METR Time Horizons","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-03-19","entrants":45,"leader":{"model":"Claude Mythos Preview (Early)","score":0.852,"effort":null,"percent":85.2,"date":"2026-04-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"Unreleased","rank":1},"rows":[{"model":"Claude Mythos Preview (Early)","score":0.852,"effort":null,"percent":85.2,"date":"2026-04-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"Unreleased","rank":1},{"model":"Claude Opus 4.6","score":0.7886,"effort":"unknown thinking","percent":78.9,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.7703,"effort":null,"percent":77.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.2","score":0.7529,"effort":"high","percent":75.3,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Opus 4.5 (16k thinking)","score":0.7497,"effort":null,"percent":75.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"GPT-5.3 Codex","score":0.7454,"effort":null,"percent":74.5,"date":"2026-02-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"GPT-5.4","score":0.7434,"effort":"xhigh","percent":74.3,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.7301,"effort":null,"percent":73.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Gemini 3 Pro Preview","score":0.7098,"effort":null,"percent":71.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.1-Codex-Max","score":0.7083,"effort":null,"percent":70.8,"date":"2025-11-19","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.1-codex-max","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1-codex-max.toml"},{"model":"GPT-5","score":0.6961,"effort":"medium","percent":69.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Sonnet 4.5 (16k thinking)","score":0.6738,"effort":null,"percent":67.4,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Claude Opus 4.1","score":0.6681,"effort":null,"percent":66.8,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Grok 4","score":0.6658,"effort":null,"percent":66.6,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"o3","score":0.6544,"effort":"medium","percent":65.4,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude Opus 4","score":0.6393,"effort":null,"percent":63.9,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"o4-mini","score":0.6393,"effort":"medium","percent":63.9,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Sonnet 4","score":0.6197,"effort":null,"percent":62.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.7 Sonnet (16k thinking)","score":0.5997,"effort":null,"percent":60.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"Kimi K2 Thinking","score":0.5919,"effort":null,"percent":59.2,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":20,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"gpt-oss-120b","score":0.5663,"effort":null,"percent":56.6,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Claude 3.7 Sonnet","score":0.5582,"effort":null,"percent":55.8,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.5544,"effort":null,"percent":55.4,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"DeepSeek-R1 (May 2025)","score":0.5378,"effort":null,"percent":53.8,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":24},{"model":"DeepSeek-R1","score":0.5193,"effort":null,"percent":51.9,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":25,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"o1","score":0.5105,"effort":"medium","percent":51.1,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"DeepSeek-V3 (Mar 2025)","score":0.4958,"effort":null,"percent":49.6,"date":"2025-03-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":27},{"model":"DeepSeek-V3","score":0.4736,"effort":null,"percent":47.4,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":28,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.4524,"effort":null,"percent":45.2,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":29},{"model":"o1-preview","score":0.4514,"effort":null,"percent":45.1,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"GPT-4o (Nov 2024)","score":0.4076,"effort":null,"percent":40.8,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.4015,"effort":null,"percent":40.2,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32},{"model":"GPT-4 Turbo (Apr 2024)","score":0.3674,"effort":null,"percent":36.7,"date":"2024-04-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33},{"model":"GPT-4 (Mar 2023)","score":0.3611,"effort":null,"percent":36.1,"date":"2023-03-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":34},{"model":"Qwen2.5-72B","score":0.3578,"effort":null,"percent":35.8,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":35},{"model":"GPT-4 Turbo Preview (January 2024)","score":0.3517,"effort":null,"percent":35.2,"date":"2024-01-25","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"GPT-4o (Aug 2024)","score":0.3384,"effort":null,"percent":33.8,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37},{"model":"Qwen2-72B","score":0.299,"effort":null,"percent":29.9,"date":"2024-06-07","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":38},{"model":"Claude 3 Opus","score":0.2948,"effort":null,"percent":29.5,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":39},{"model":"GPT-4 (Jun 2023)","score":0.293,"effort":null,"percent":29.3,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"GPT-4 Turbo Preview (Nov 2023)","score":0.289,"effort":null,"percent":28.9,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"GPT-4 Turbo (Nov 2023)","score":0.2718,"effort":null,"percent":27.2,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":42},{"model":"GPT-3.5 Turbo Instruct","score":0.2146,"effort":null,"percent":21.5,"date":"2023-09-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43},{"model":"GPT-3 175B (davinci)","score":0.1619,"effort":null,"percent":16.2,"date":null,"organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":44},{"model":"GPT-2 (1.5B)","score":0.101,"effort":null,"percent":10.1,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":45}]},{"id":"hle","name":"HLE","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.048,"released":"2025-01-23","entrants":43,"leader":{"model":"Claude Fable 5.1","score":0.465,"effort":"xhigh","percent":46.5,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},"rows":[{"model":"Claude Fable 5.1","score":0.465,"effort":"xhigh","percent":46.5,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.4644,"effort":null,"percent":46.4,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.4 Pro","score":0.4432,"effort":null,"percent":44.3,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.4-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-pro.toml"},{"model":"Muse Spark","score":0.4056,"effort":null,"percent":40.6,"date":"2026-04-08","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Gemini 3 Pro Preview","score":0.3752,"effort":null,"percent":37.5,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.4","score":0.3624,"effort":"xhigh","percent":36.2,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Claude Opus 4.7","score":0.362,"effort":"unknown","percent":36.2,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Claude Opus 4.6","score":0.3444,"effort":"max","percent":34.4,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5 Pro","score":0.3164,"effort":null,"percent":31.6,"date":"2025-10-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-pro.toml"},{"model":"GPT-5.2","score":0.278,"effort":"unknown thinking","percent":27.8,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"GPT-5","score":0.2532,"effort":"high","percent":25.3,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Opus 4.5","score":0.252,"effort":"unknown thinking","percent":25.2,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Kimi K2.5","score":0.2437,"effort":null,"percent":24.4,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":13},{"model":"GPT-5.1","score":0.2368,"effort":"unknown thinking","percent":23.7,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.2164,"effort":null,"percent":21.6,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"o3","score":0.2032,"effort":"high","percent":20.3,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"GPT-5 mini","score":0.1944,"effort":"unknown thinking","percent":19.4,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.19,"effort":null,"percent":19.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.1816,"effort":null,"percent":18.2,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"o4-mini","score":0.1808,"effort":"high","percent":18.1,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Claude Sonnet 4.5","score":0.1372,"effort":"unknown thinking","percent":13.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":0.1208,"effort":null,"percent":12.1,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":22},{"model":"Claude Opus 4.1","score":0.1152,"effort":"unknown thinking","percent":11.5,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Gemini 2.5 Flash (May 2025)","score":0.1096,"effort":null,"percent":11.0,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":24},{"model":"Claude Opus 4","score":0.1072,"effort":null,"percent":10.7,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.0864,"effort":null,"percent":8.6,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"GLM-4.5","score":0.0832,"effort":null,"percent":8.3,"date":"2025-08-03","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27,"model_id":"zhipuai/glm-4.5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.5.toml"},{"model":"GLM-4.5-Air","score":0.0812,"effort":null,"percent":8.1,"date":"2025-07-20","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":28,"model_id":"zhipuai/glm-4.5-air","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.5-air.toml"},{"model":"o1 Pro","score":0.0812,"effort":null,"percent":8.1,"date":"2025-03-19","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":29,"model_id":"openai/o1-pro","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1-pro.toml"},{"model":"Claude 3.7 Sonnet","score":0.0804,"effort":"unknown thinking","percent":8.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"o1","score":0.0796,"effort":"unknown thinking","percent":8.0,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude Sonnet 4","score":0.0776,"effort":"unknown thinking","percent":7.8,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":32,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"GPT-5.1 (no thinking)","score":0.068,"effort":null,"percent":6.8,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":33},{"model":"Gemini 2.0 Flash Thinking Exp","score":0.0656,"effort":null,"percent":6.6,"date":"2025-01-21","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":34},{"model":"Llama 4 Maverick","score":0.0568,"effort":null,"percent":5.7,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":35},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.0544,"effort":null,"percent":5.4,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":36},{"model":"GPT-4.1","score":0.054,"effort":null,"percent":5.4,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":37,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.046,"effort":null,"percent":4.6,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":38},{"model":"Mistral Medium 3","score":0.0452,"effort":null,"percent":4.5,"date":"2025-05-07","organization":"Mistral AI","countries":["France"],"accessibility":"API access","rank":39,"model_id":"mistral/mistral-medium-2505","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2505.toml"},{"model":"Amazon Nova Pro","score":0.044,"effort":null,"percent":4.4,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":40},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.0408,"effort":null,"percent":4.1,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":41},{"model":"Amazon Nova Lite","score":0.0364,"effort":null,"percent":3.6,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":42},{"model":"GPT-4o (Nov 2024)","score":0.0272,"effort":null,"percent":2.7,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":43}]},{"id":"openbookqa","name":"OpenBookQA","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2018-09-08","entrants":42,"leader":{"model":"phi-3-mini 3.8B","score":0.88,"effort":null,"percent":88.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},"rows":[{"model":"phi-3-mini 3.8B","score":0.88,"effort":null,"percent":88.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},{"model":"phi-3-small 7.4B","score":0.88,"effort":null,"percent":88.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":2},{"model":"phi-3-medium 14B","score":0.874,"effort":null,"percent":87.4,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":3},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.86,"effort":null,"percent":86.0,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Mixtral 8x7B","score":0.858,"effort":null,"percent":85.8,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":5},{"model":"Llama 3-8B","score":0.826,"effort":null,"percent":82.6,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":6},{"model":"Mistral 7B v0.1","score":0.798,"effort":null,"percent":79.8,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"Gemma 7B","score":0.786,"effort":null,"percent":78.6,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":8},{"model":"Phi-2","score":0.736,"effort":null,"percent":73.6,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"PaLM (540B)","score":0.68,"effort":null,"percent":68.0,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":10},{"model":"InstructGPT 175B","score":0.654,"effort":null,"percent":65.4,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"Falcon-180B","score":0.642,"effort":null,"percent":64.2,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":12},{"model":"GLaM","score":0.63,"effort":null,"percent":63.0,"date":"2021-12-13","organization":"Google","countries":["United States of America"],"accessibility":"Unreleased","rank":13},{"model":"Llama 2-70B","score":0.602,"effort":null,"percent":60.2,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":14},{"model":"LLaMA-65B","score":0.602,"effort":null,"percent":60.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":15},{"model":"Llama 2-7B","score":0.586,"effort":null,"percent":58.6,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":16},{"model":"LLaMA-33B","score":0.586,"effort":null,"percent":58.6,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":17},{"model":"Llama 2-34B","score":0.582,"effort":null,"percent":58.2,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":18},{"model":"PaLM 2-M","score":0.574,"effort":null,"percent":57.4,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":19},{"model":"LLaMA-7B","score":0.572,"effort":null,"percent":57.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":20},{"model":"Llama 2-13B","score":0.57,"effort":null,"percent":57.0,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"Falcon-40B","score":0.566,"effort":null,"percent":56.6,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"LLaMA-13B","score":0.564,"effort":null,"percent":56.4,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":23},{"model":"PaLM 2-S","score":0.562,"effort":null,"percent":56.2,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":24},{"model":"MPT-30B","score":0.52,"effort":null,"percent":52.0,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":25},{"model":"Falcon-7B","score":0.516,"effort":null,"percent":51.6,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":26},{"model":"MPT-7B","score":0.514,"effort":null,"percent":51.4,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":27},{"model":"PaLM 62B","score":0.504,"effort":null,"percent":50.4,"date":"2022-04-04","organization":null,"countries":[],"accessibility":null,"rank":28},{"model":"XGen-7B","score":0.402,"effort":null,"percent":40.2,"date":"2023-06-27","organization":"Salesforce","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":29},{"model":"RedPajama-INCITE-7B-Base","score":0.4,"effort":null,"percent":40.0,"date":"2023-05-04","organization":null,"countries":[],"accessibility":null,"rank":30},{"model":"opt-13b","score":0.398,"effort":null,"percent":39.8,"date":"2022-05-11","organization":null,"countries":[],"accessibility":null,"rank":31},{"model":"Dolly 2.0-12b","score":0.392,"effort":null,"percent":39.2,"date":"2023-04-11","organization":"Databricks","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"open_llama_7b","score":0.39,"effort":null,"percent":39.0,"date":"2023-06-07","organization":null,"countries":[],"accessibility":null,"rank":33},{"model":"GPT-NeoX-20B","score":0.388,"effort":null,"percent":38.8,"date":"2022-04-07","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"GPT-J-6B","score":0.382,"effort":null,"percent":38.2,"date":"2021-08-05","organization":"EleutherAI,LAION","countries":["Germany","United States of America"],"accessibility":"Open weights (unrestricted)","rank":35},{"model":"Phi-1.5","score":0.372,"effort":null,"percent":37.2,"date":"2023-09-11","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":36},{"model":"Cerebras-GPT-13B","score":0.358,"effort":null,"percent":35.8,"date":"2023-03-20","organization":"Cerebras Systems","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":37},{"model":"vicuna-13b-v1.1","score":0.33,"effort":null,"percent":33.0,"date":"2023-04-12","organization":null,"countries":[],"accessibility":null,"rank":38},{"model":"stablelm-tuned-alpha-7b","score":0.324,"effort":null,"percent":32.4,"date":"2023-04-19","organization":null,"countries":[],"accessibility":null,"rank":39},{"model":"OPT-1.3B","score":0.24,"effort":null,"percent":24.0,"date":"2022-05-11","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":40},{"model":"GPT-Neo-2.7B","score":0.232,"effort":null,"percent":23.2,"date":"2021-03-30","organization":"EleutherAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":41},{"model":"GPT-2 (1.5B)","score":0.224,"effort":null,"percent":22.4,"date":"2019-11-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":42}]},{"id":"triviaqa","name":"TriviaQA","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2017-05-09","entrants":38,"leader":{"model":"Llama 2-70B","score":0.876,"effort":null,"percent":87.6,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":1},"rows":[{"model":"Llama 2-70B","score":0.876,"effort":null,"percent":87.6,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":1},{"model":"Claude 2","score":0.875,"effort":null,"percent":87.5,"date":"2023-07-11","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Claude 1.3","score":0.867,"effort":null,"percent":86.7,"date":"2023-04-18","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"PaLM 2-L","score":0.861,"effort":null,"percent":86.1,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":4},{"model":"LLaMA-65B","score":0.86,"effort":null,"percent":86.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":5},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.858,"effort":null,"percent":85.8,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"GPT-4 (Jun 2023)","score":0.848,"effort":null,"percent":84.8,"date":"2023-06-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Llama 2-34B","score":0.846,"effort":null,"percent":84.6,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Unreleased","rank":8},{"model":"LLaMA-33B","score":0.838,"effort":null,"percent":83.8,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":9},{"model":"DeepSeek-V3","score":0.829,"effort":null,"percent":82.9,"date":"2024-12-26","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":10,"model_id":"deepseek/deepseek-v3","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v3.toml"},{"model":"Llama 3.1-405B","score":0.827,"effort":null,"percent":82.7,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":11},{"model":"Mixtral 8x7B","score":0.822,"effort":null,"percent":82.2,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":12},{"model":"PaLM 2-M","score":0.817,"effort":null,"percent":81.7,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":13},{"model":"PaLM (540B)","score":0.814,"effort":null,"percent":81.4,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":14},{"model":"DeepSeek-V2 (MoE-236B, May 2024)","score":0.8,"effort":null,"percent":80.0,"date":"2024-05-07","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (restricted use)","rank":15},{"model":"Falcon-40B","score":0.799,"effort":null,"percent":79.9,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":16},{"model":"Llama 2-13B","score":0.796,"effort":null,"percent":79.6,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":17},{"model":"Claude Instant","score":0.789,"effort":null,"percent":78.9,"date":null,"organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"LLaMA-13B","score":0.779,"effort":null,"percent":77.9,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":19},{"model":"GLaM","score":0.758,"effort":null,"percent":75.8,"date":"2021-12-13","organization":"Google","countries":["United States of America"],"accessibility":"Unreleased","rank":20},{"model":"PaLM 2-S","score":0.752,"effort":null,"percent":75.2,"date":"2023-05-17","organization":null,"countries":[],"accessibility":null,"rank":21},{"model":"Mistral 7B v0.1","score":0.752,"effort":null,"percent":75.2,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"phi-3-medium 14B","score":0.739,"effort":null,"percent":73.9,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":23},{"model":"Llama 2-7B","score":0.737,"effort":null,"percent":73.7,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":24},{"model":"MPT-30B","score":0.736,"effort":null,"percent":73.6,"date":"2023-06-22","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":25},{"model":"Gemma 7B","score":0.723,"effort":null,"percent":72.3,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":26},{"model":"Qwen2.5-72B","score":0.719,"effort":null,"percent":71.9,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":27},{"model":"InstructGPT 175B","score":0.712,"effort":null,"percent":71.2,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"LLaMA-7B","score":0.71,"effort":null,"percent":71.0,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":29},{"model":"Llama 3-8B","score":0.677,"effort":null,"percent":67.7,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":30},{"model":"Falcon-7B","score":0.646,"effort":null,"percent":64.6,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":31},{"model":"Chinchilla","score":0.646,"effort":null,"percent":64.6,"date":"2022-03-29","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":32},{"model":"phi-3-mini 3.8B","score":0.64,"effort":null,"percent":64.0,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":33},{"model":"MPT-7B","score":0.616,"effort":null,"percent":61.6,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":34},{"model":"phi-3-small 7.4B","score":0.581,"effort":null,"percent":58.1,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":35},{"model":"Gopher (280B)","score":0.572,"effort":null,"percent":57.2,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":36},{"model":"Gemma 2B","score":0.532,"effort":null,"percent":53.2,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":37},{"model":"Phi-2","score":0.452,"effort":null,"percent":45.2,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":38}]},{"id":"balrog","name":"Balrog","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2024-11-20","entrants":34,"leader":{"model":"Gemini 3 Pro Preview","score":0.581,"effort":null,"percent":58.1,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},"rows":[{"model":"Gemini 3 Pro Preview","score":0.581,"effort":null,"percent":58.1,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.57,"effort":null,"percent":57.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Gemini 3 Flash Preview","score":0.481,"effort":null,"percent":48.1,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Grok 4","score":0.436,"effort":null,"percent":43.6,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Claude Opus 4.5 (no thinking)","score":0.435,"effort":null,"percent":43.5,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.433,"effort":null,"percent":43.3,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"Claude Opus 4.5 (64k thinking)","score":0.43,"effort":null,"percent":43.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Claude Opus 4.5","score":0.43,"effort":"unknown thinking","percent":43.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"DeepSeek-R1","score":0.349,"effort":null,"percent":34.9,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":9,"model_id":"deepseek/deepseek-r1","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1.toml"},{"model":"Gemini 2.5 Flash (Jun 2025)","score":0.335,"effort":null,"percent":33.5,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"GPT-5","score":0.328,"effort":"minimal","percent":32.8,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.326,"effort":null,"percent":32.6,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"GPT-4o (May 2024)","score":0.323,"effort":null,"percent":32.3,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"Claude Haiku 4.5","score":0.312,"effort":null,"percent":31.2,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"Grok 3","score":0.295,"effort":null,"percent":29.5,"date":"2025-04-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Reka Flash 3","score":0.292,"effort":null,"percent":29.2,"date":"2025-03-10","organization":"Reka AI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":16},{"model":"Llama 3.1-70B","score":0.279,"effort":null,"percent":27.9,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":17},{"model":"Llama 3.2 90B","score":0.273,"effort":null,"percent":27.3,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":18},{"model":"Llama 3.3 70B","score":0.23,"effort":null,"percent":23.0,"date":"2024-12-06","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":19},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.21,"effort":null,"percent":21.0,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"DeepSeek-R1-Distill-Qwen-32B","score":0.195,"effort":null,"percent":19.5,"date":"2025-01-20","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"deepseek/deepseek-r1-distill-qwen-32b","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-r1-distill-qwen-32b.toml"},{"model":"Claude 3.5 Haiku (Oct 2024)","score":0.193,"effort":null,"percent":19.3,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":22},{"model":"Mistral NeMo","score":0.176,"effort":null,"percent":17.6,"date":"2024-07-18","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":23,"model_id":"mistral/mistral-nemo","lab":"mistral","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-nemo.toml"},{"model":"GPT-4o mini","score":0.174,"effort":null,"percent":17.4,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Llama 3.2 11B","score":0.168,"effort":null,"percent":16.8,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":25},{"model":"Qwen2.5-72B","score":0.162,"effort":null,"percent":16.2,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26},{"model":"Llama 3.1-8B","score":0.151,"effort":null,"percent":15.1,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":27},{"model":"Gemini 1.5 Flash (Sep 2024)","score":0.146,"effort":null,"percent":14.6,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Qwen2-VL-72B-Instruct","score":0.128,"effort":null,"percent":12.8,"date":"2024-08-29","organization":null,"countries":[],"accessibility":null,"rank":29},{"model":"Phi-4","score":0.116,"effort":null,"percent":11.6,"date":"2024-12-12","organization":"Microsoft Research","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":30},{"model":"Llama 3.2 3B","score":0.101,"effort":null,"percent":10.1,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":31,"model_id":"meta/llama-3.2-3b","lab":"meta","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/llama-3.2-3b.toml"},{"model":"Qwen2.5-7B","score":0.078,"effort":null,"percent":7.8,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":32},{"model":"Llama 3.2 1B","score":0.066,"effort":null,"percent":6.6,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":33,"model_id":"meta/llama-3.2-1b","lab":"meta","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/llama-3.2-1b.toml"},{"model":"Qwen2-VL-7B-Instruct","score":0.037,"effort":null,"percent":3.7,"date":"2024-08-29","organization":null,"countries":[],"accessibility":null,"rank":34}]},{"id":"swe-bench-verified","name":"SWE-Bench verified","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2024-08-13","entrants":32,"leader":{"model":"Claude Opus 4.7","score":0.8347,"effort":"max","percent":83.5,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},"rows":[{"model":"Claude Opus 4.7","score":0.8347,"effort":"max","percent":83.5,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.5","score":0.8058,"effort":"xhigh","percent":80.6,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.5 Flash","score":0.7934,"effort":null,"percent":79.3,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.7872,"effort":null,"percent":78.7,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"GLM-5.2","score":0.787,"effort":null,"percent":78.7,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":5,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"DeepSeek v4 Pro","score":0.7764,"effort":"max","percent":77.6,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Qwen3.7 Max","score":0.7727,"effort":null,"percent":77.3,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":7,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"GPT-5.4","score":0.7686,"effort":"high","percent":76.9,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Qwen 3.6 Max (Preview)","score":0.7665,"effort":null,"percent":76.7,"date":"2026-04-20","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":9},{"model":"Kimi K2.6","score":0.7665,"effort":null,"percent":76.7,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":10,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.7665,"effort":null,"percent":76.7,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"Gemini 3.1 Pro Preview","score":0.7562,"effort":null,"percent":75.6,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Gemini 3 Flash Preview","score":0.7541,"effort":null,"percent":75.4,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Claude Sonnet 4.6 (no thinking)","score":0.7521,"effort":null,"percent":75.2,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"GPT-5.3 Codex","score":0.7479,"effort":"high","percent":74.8,"date":"2026-02-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5.3-codex","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.3-codex.toml"},{"model":"GLM-5.1","score":0.7417,"effort":null,"percent":74.2,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Kimi K2.5","score":0.7376,"effort":null,"percent":73.8,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17},{"model":"GPT-5.2","score":0.7376,"effort":"high","percent":73.8,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"GPT-5","score":0.7355,"effort":"high","percent":73.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Opus 4.1","score":0.7335,"effort":null,"percent":73.3,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Gemini 3 Pro Preview","score":0.7293,"effort":null,"percent":72.9,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GLM-5","score":0.7208,"effort":null,"percent":72.1,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.7128,"effort":null,"percent":71.3,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"Claude Opus 4","score":0.7066,"effort":null,"percent":70.7,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"GPT-5.1","score":0.6798,"effort":"high","percent":68.0,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"GPT-5 mini","score":0.6467,"effort":"medium","percent":64.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"o3","score":0.6232,"effort":"medium","percent":62.3,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":27,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude 3.7 Sonnet","score":0.6095,"effort":null,"percent":61.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.5785,"effort":null,"percent":57.9,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":29},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.5756,"effort":null,"percent":57.6,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":30},{"model":"GPT-4.1","score":0.4854,"effort":null,"percent":48.5,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"GPT-4o (Nov 2024)","score":0.3099,"effort":null,"percent":31.0,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":32}]},{"id":"gso-bench","name":"GSO-Bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-05-29","entrants":31,"leader":{"model":"Claude Opus 4.8","score":0.4706,"effort":null,"percent":47.1,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},"rows":[{"model":"Claude Opus 4.8","score":0.4706,"effort":null,"percent":47.1,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Claude Opus 4.7","score":0.4412,"effort":"high","percent":44.1,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Claude Opus 4.7 (no thinking)","score":0.4412,"effort":null,"percent":44.1,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Claude Opus 4.6","score":0.412,"effort":"high","percent":41.2,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5.5","score":0.402,"effort":"xhigh","percent":40.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Sonnet 5","score":0.3725,"effort":"unknown thinking","percent":37.2,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Claude Opus 4.6 (no thinking)","score":0.3333,"effort":null,"percent":33.3,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"GPT-5.4","score":0.3137,"effort":"xhigh","percent":31.4,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GPT-5.2","score":0.2745,"effort":"high","percent":27.5,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.265,"effort":null,"percent":26.5,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"Claude Opus 4.5","score":0.2647,"effort":"unknown thinking","percent":26.5,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.2255,"effort":null,"percent":22.6,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Gemini 3 Pro Preview","score":0.1863,"effort":null,"percent":18.6,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Claude Sonnet 4.5","score":0.1471,"effort":"unknown thinking","percent":14.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.147,"effort":null,"percent":14.7,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"GPT-5.1","score":0.1373,"effort":"unknown thinking","percent":13.7,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Gemini 3 Flash Preview","score":0.098,"effort":null,"percent":9.8,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"o3","score":0.088,"effort":"high","percent":8.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude Opus 4","score":0.069,"effort":null,"percent":6.9,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"GPT-5","score":0.0686,"effort":"high","percent":6.9,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Sonnet 4","score":0.049,"effort":null,"percent":4.9,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Kimi K2 Instruct","score":0.049,"effort":null,"percent":4.9,"date":"2025-07-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"Qwen3-Coder-480B-A35B","score":0.049,"effort":null,"percent":4.9,"date":"2025-07-31","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":23},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.046,"effort":null,"percent":4.6,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.0392,"effort":null,"percent":3.9,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.0392,"effort":null,"percent":3.9,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":26},{"model":"Claude 3.7 Sonnet","score":0.038,"effort":null,"percent":3.8,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27},{"model":"o4-mini","score":0.036,"effort":"high","percent":3.6,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":28,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"GLM-4.5-Air","score":0.0294,"effort":null,"percent":2.9,"date":"2025-07-20","organization":"Z.ai (Zhipu AI),Tsinghua University","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":29,"model_id":"zhipuai/glm-4.5-air","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-4.5-air.toml"},{"model":"o3-mini","score":0.013,"effort":null,"percent":1.3,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":30,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"GPT-4o (Nov 2024)","score":0.0,"effort":null,"percent":0.0,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":31}]},{"id":"frontiercode","name":"FrontierCode","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-08","entrants":30,"leader":{"model":"Claude Fable 5","score":0.5348,"effort":"unknown","percent":53.5,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.5348,"effort":"unknown","percent":53.5,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 5","score":0.5338,"effort":"max","percent":53.4,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-6 Astra","score":0.5326,"effort":"max","percent":53.3,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5.1","score":0.5091,"effort":"medium","percent":50.9,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Grok 4.6","score":0.4801,"effort":"unknown","percent":48.0,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.6 Sol","score":0.4749,"effort":"unknown thinking","percent":47.5,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 4.8","score":0.465,"effort":null,"percent":46.5,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Kimi K3","score":0.4417,"effort":"unknown","percent":44.2,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":8,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Gemini 3.7 Flash","score":0.4359,"effort":"unknown","percent":43.6,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GPT-5.5","score":0.4296,"effort":"unknown thinking","percent":43.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Sonnet 5","score":0.4273,"effort":"unknown thinking","percent":42.7,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Grok 4.5","score":0.4244,"effort":"unknown thinking","percent":42.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"SWE-1.7","score":0.4199,"effort":null,"percent":42.0,"date":"2026-07-08","organization":"Cognition","countries":["United States of America"],"accessibility":null,"rank":13},{"model":"GPT-5.6 Terra","score":0.4131,"effort":"unknown thinking","percent":41.3,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GPT-5.6 Luna","score":0.3981,"effort":"unknown thinking","percent":39.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Opus 4.7","score":0.3854,"effort":"unknown","percent":38.5,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Gemini 3.6 Flash","score":0.3437,"effort":null,"percent":34.4,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"Kimi K2.7 Code","score":0.3006,"effort":null,"percent":30.1,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"GPT-5.4 mini","score":0.2704,"effort":"unknown thinking","percent":27.0,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"Claude Opus 4.6","score":0.2664,"effort":"unknown thinking","percent":26.6,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Composer 2.5","score":0.2564,"effort":null,"percent":25.6,"date":"2026-05-18","organization":"Cursor","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"GLM-5.2","score":0.245,"effort":null,"percent":24.5,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Claude Sonnet 4.6","score":0.2431,"effort":"unknown thinking","percent":24.3,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"DeepSeek V4 Flash 0731","score":0.188,"effort":"unknown","percent":18.8,"date":"2026-07-31","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":24,"model_id":"deepseek/deepseek-v4-flash-0731","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-flash-0731.toml"},{"model":"DeepSeek v4 Pro","score":0.1764,"effort":"unknown thinking","percent":17.6,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":25,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"MiniMax-M3","score":0.147,"effort":null,"percent":14.7,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":26},{"model":"Inkling","score":0.14,"effort":null,"percent":14.0,"date":"2026-07-15","organization":"Thinking Machines","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":27},{"model":"Qwen3.7 Plus","score":0.102,"effort":null,"percent":10.2,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":28,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"},{"model":"SWE-1.6","score":0.0939,"effort":null,"percent":9.4,"date":"2026-04-07","organization":"Cognition","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":29},{"model":"Mistral Medium 3.5","score":0.08,"effort":null,"percent":8.0,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":30,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"}]},{"id":"vpct","name":"VPCT","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.3333333333333333,"released":"2025-01-30","entrants":29,"leader":{"model":"Gemini 3 Pro Preview","score":0.91,"effort":null,"percent":91.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},"rows":[{"model":"Gemini 3 Pro Preview","score":0.91,"effort":null,"percent":91.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5.2","score":0.84,"effort":"xhigh","percent":84.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3 Flash Preview","score":0.726,"effort":null,"percent":72.6,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"GPT-5","score":0.66,"effort":"high","percent":66.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GPT-5.1","score":0.587,"effort":"high","percent":58.7,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"o4-mini","score":0.575,"effort":"medium","percent":57.5,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"o3","score":0.52,"effort":"medium","percent":52.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.48,"effort":null,"percent":48.0,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.464,"effort":null,"percent":46.4,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Gemini 2.5 Flash (Sep 2025)","score":0.462,"effort":null,"percent":46.2,"date":"2025-09-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.45,"effort":null,"percent":45.0,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"gemini-robotics-er-1.5-preview","score":0.408,"effort":null,"percent":40.8,"date":"2025-09-26","organization":null,"countries":[],"accessibility":null,"rank":12},{"model":"GPT-5 mini","score":0.402,"effort":"medium","percent":40.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-5-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-mini.toml"},{"model":"GPT-4o (Nov 2024)","score":0.4,"effort":null,"percent":40.0,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"Claude Opus 4.5 (32k thinking)","score":0.4,"effort":null,"percent":40.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Claude Sonnet 4.5 (32k thinking)","score":0.398,"effort":null,"percent":39.8,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"Claude 3.7 Sonnet","score":0.39,"effort":null,"percent":39.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":0.38,"effort":null,"percent":38.0,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"Claude Opus 4","score":0.38,"effort":null,"percent":38.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.38,"effort":null,"percent":38.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"GPT-5 nano","score":0.372,"effort":"high","percent":37.2,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"openai/gpt-5-nano","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5-nano.toml"},{"model":"o1","score":0.37,"effort":"medium","percent":37.0,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude 3.7 Sonnet (64k thinking)","score":0.35,"effort":null,"percent":35.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":23},{"model":"Claude Opus 4.1","score":0.35,"effort":null,"percent":35.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"GPT-4o mini","score":0.34,"effort":null,"percent":34.0,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Claude Sonnet 4","score":0.34,"effort":null,"percent":34.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.33,"effort":null,"percent":33.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":27},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.33,"effort":null,"percent":33.0,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28},{"model":"Gemini 2.5 Flash-Lite (Sep 2025)","score":0.3,"effort":null,"percent":30.0,"date":"2025-09-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":29}]},{"id":"geobench","name":"GeoBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-03-01","entrants":28,"leader":{"model":"Gemini 3 Flash Preview","score":0.88,"effort":null,"percent":88.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},"rows":[{"model":"Gemini 3 Flash Preview","score":0.88,"effort":null,"percent":88.0,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.86,"effort":null,"percent":86.0,"date":"2025-05-06","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Gemini 3 Pro Preview","score":0.84,"effort":null,"percent":84.0,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Gemini 2.5 Pro Exp (Mar 2025)","score":0.81,"effort":null,"percent":81.0,"date":"2025-03-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"GPT-5","score":0.81,"effort":"medium","percent":81.0,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o1","score":0.8,"effort":"medium","percent":80.0,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.77,"effort":null,"percent":77.0,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Gemini 1.5 Flash (Sep 2024)","score":0.76,"effort":null,"percent":76.0,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Gemini 2.5 Flash (May 2025)","score":0.76,"effort":null,"percent":76.0,"date":"2025-05-20","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Claude Opus 4.5 (no thinking)","score":0.75,"effort":null,"percent":75.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"o3","score":0.74,"effort":"medium","percent":74.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemini 2.5 Flash Preview (Apr 2025)","score":0.73,"effort":null,"percent":73.0,"date":"2025-04-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"GPT-4.1","score":0.72,"effort":null,"percent":72.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"GPT-4o (Nov 2024)","score":0.71,"effort":null,"percent":71.0,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"Claude 3.7 Sonnet","score":0.68,"effort":null,"percent":68.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Claude 3.7 Sonnet (15k thinking)","score":0.65,"effort":null,"percent":65.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"o4-mini","score":0.64,"effort":"high","percent":64.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"GPT-4o mini","score":0.64,"effort":null,"percent":64.0,"date":"2024-07-18","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"openai/gpt-4o-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4o-mini.toml"},{"model":"Qwen2.5-72B","score":0.62,"effort":null,"percent":62.0,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.62,"effort":null,"percent":62.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":20},{"model":"Gemma 3 27B","score":0.52,"effort":null,"percent":52.0,"date":"2025-03-12","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"Llama 4 Maverick","score":0.52,"effort":null,"percent":52.0,"date":"2025-04-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"Llama 3.2 90B","score":0.52,"effort":null,"percent":52.0,"date":"2024-09-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":23},{"model":"Claude Opus 4","score":0.49,"effort":null,"percent":49.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Grok 4","score":0.45,"effort":null,"percent":45.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":25},{"model":"Claude Sonnet 4","score":0.37,"effort":null,"percent":37.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Pixtral 12B","score":0.37,"effort":null,"percent":37.0,"date":"2024-09-17","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":27,"model_id":"mistral/pixtral-12b","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/pixtral-12b.toml"},{"model":"Claude 3.5 Haiku (Oct 2024)","score":0.34,"effort":null,"percent":34.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":28}]},{"id":"deepresearch-bench","name":"DeepResearch Bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-06-13","entrants":26,"leader":{"model":"Claude Opus 4.6","score":0.553,"effort":"high","percent":55.3,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},"rows":[{"model":"Claude Opus 4.6","score":0.553,"effort":"high","percent":55.3,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Claude Sonnet 4.6","score":0.549,"effort":"high","percent":54.9,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Claude Opus 4.5","score":0.548,"effort":"high","percent":54.8,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"GPT-5.5","score":0.54,"effort":"high","percent":54.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Sonnet 4.5 (2k thinking)","score":0.526,"effort":null,"percent":52.6,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Claude Opus 4.8","score":0.502,"effort":null,"percent":50.2,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Gemini 3 Flash Preview","score":0.498,"effort":null,"percent":49.8,"date":"2025-12-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"google/gemini-3-flash-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-flash-preview.toml"},{"model":"GPT-5","score":0.496,"effort":"low","percent":49.6,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Claude Opus 4.1","score":0.483,"effort":null,"percent":48.3,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.478,"effort":null,"percent":47.8,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Sonnet 4.5","score":0.475,"effort":"low","percent":47.5,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Grok 4","score":0.473,"effort":null,"percent":47.3,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Claude Opus 4","score":0.468,"effort":null,"percent":46.8,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Claude Sonnet 4","score":0.466,"effort":null,"percent":46.6,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Gemini 3 Pro Preview","score":0.463,"effort":"low","percent":46.3,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Claude Haiku 4.5","score":0.455,"effort":"low","percent":45.5,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"o3","score":0.452,"effort":"medium","percent":45.2,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Claude 3.7 Sonnet (2k thinking)","score":0.436,"effort":null,"percent":43.6,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.428,"effort":null,"percent":42.8,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19},{"model":"GPT-5.1","score":0.428,"effort":"low","percent":42.8,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.415,"effort":null,"percent":41.5,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"GPT-5.2","score":0.411,"effort":"low","percent":41.1,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":22,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Gemini 3.1 Flash-Lite","score":0.373,"effort":null,"percent":37.3,"date":"2026-03-03","organization":"Google","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"google/gemini-3.1-flash-lite","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-flash-lite.toml"},{"model":"GPT-5.4 mini","score":0.363,"effort":"low","percent":36.3,"date":"2026-03-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":24,"model_id":"openai/gpt-5.4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4-mini.toml"},{"model":"GPT-5.4","score":0.351,"effort":"low","percent":35.1,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"DeepSeek-R1 (May 2025)","score":0.351,"effort":null,"percent":35.1,"date":"2025-05-28","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":26}]},{"id":"lambada","name":"LAMBADA","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2016-06-20","entrants":26,"leader":{"model":"Megatron-Turing NLG 530B","score":0.8715,"effort":null,"percent":87.2,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":1},"rows":[{"model":"Megatron-Turing NLG 530B","score":0.8715,"effort":null,"percent":87.2,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":1},{"model":"InstructGPT 175B","score":0.864,"effort":null,"percent":86.4,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Falcon-180B","score":0.798,"effort":null,"percent":79.8,"date":"2023-09-06","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (restricted use)","rank":3},{"model":"Llama 2-70B","score":0.789,"effort":null,"percent":78.9,"date":"2023-07-18","organization":"Meta AI","countries":[],"accessibility":null,"rank":4},{"model":"Inflection-1","score":0.785,"effort":null,"percent":78.5,"date":"2023-06-22","organization":"Inflection AI","countries":["United States of America"],"accessibility":"Hosted access (no API)","rank":5},{"model":"PaLM (540B)","score":0.779,"effort":null,"percent":77.9,"date":"2022-04-04","organization":"Google Research","countries":["United States of America"],"accessibility":"Unreleased","rank":6},{"model":"LLaMA-65B","score":0.777,"effort":null,"percent":77.7,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":7},{"model":"Chinchilla","score":0.774,"effort":null,"percent":77.4,"date":"2022-03-29","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":8},{"model":"Falcon-40B","score":0.773,"effort":null,"percent":77.3,"date":"2023-03-15","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"LLaMA-33B","score":0.772,"effort":null,"percent":77.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":10},{"model":"Llama 2-13B","score":0.765,"effort":null,"percent":76.5,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":11},{"model":"LLaMA-13B","score":0.752,"effort":null,"percent":75.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":12},{"model":"Falcon-7B","score":0.749,"effort":null,"percent":74.9,"date":"2023-04-24","organization":"Technology Innovation Institute","countries":["United Arab Emirates"],"accessibility":"Open weights (unrestricted)","rank":13},{"model":"Gopher (280B)","score":0.745,"effort":null,"percent":74.5,"date":"2021-12-08","organization":"DeepMind","countries":["United Kingdom"],"accessibility":"Unreleased","rank":14},{"model":"Baichuan2-13B","score":0.74,"effort":null,"percent":74.0,"date":"2023-09-06","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":15},{"model":"Baichuan 2-7B","score":0.733,"effort":null,"percent":73.3,"date":"2023-09-20","organization":"Baichuan","countries":["China"],"accessibility":"Open weights (restricted use)","rank":16},{"model":"LLaMA-7B","score":0.733,"effort":null,"percent":73.3,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":17},{"model":"Llama 2-7B","score":0.733,"effort":null,"percent":73.3,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":18},{"model":"internlm-20b","score":0.718,"effort":null,"percent":71.8,"date":"2023-09-18","organization":null,"countries":[],"accessibility":null,"rank":19},{"model":"Stable Beluga 2","score":0.713,"effort":null,"percent":71.3,"date":"2023-07-20","organization":"Stability AI","countries":["United Kingdom"],"accessibility":"Open weights (non-commercial)","rank":20},{"model":"Qwen-14B","score":0.711,"effort":null,"percent":71.1,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":21},{"model":"MPT-7B","score":0.7,"effort":null,"percent":70.0,"date":"2023-05-05","organization":"MosaicML","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":22},{"model":"Qwen-7B","score":0.679,"effort":null,"percent":67.9,"date":"2023-09-28","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (restricted use)","rank":23},{"model":"internlm-7b","score":0.67,"effort":null,"percent":67.0,"date":"2023-07-05","organization":null,"countries":[],"accessibility":null,"rank":24},{"model":"Qwen-1.8B","score":0.584,"effort":null,"percent":58.4,"date":"2023-11-30","organization":null,"countries":[],"accessibility":null,"rank":25},{"model":"chatglm2-6b","score":0.543,"effort":null,"percent":54.3,"date":"2023-06-24","organization":null,"countries":[],"accessibility":null,"rank":26}]},{"id":"deepswe","name":"DeepSWE","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-05-26","entrants":26,"leader":{"model":"GPT-6 Astra","score":0.7412,"effort":"xhigh","percent":74.1,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.7412,"effort":"xhigh","percent":74.1,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Gemini 3.8 Flash","score":0.7383,"effort":"high","percent":73.8,"date":"2026-09-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"google/gemini-3.8-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.8-flash.toml"},{"model":"Claude Opus 5","score":0.7365,"effort":"max","percent":73.6,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.6 Sol","score":0.7267,"effort":"max","percent":72.7,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Fable 5","score":0.6991,"effort":"xhigh","percent":69.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.6 Terra","score":0.6962,"effort":"max","percent":69.6,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"GLM-5.3","score":0.6896,"effort":"max","percent":69.0,"date":"2026-08-14","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"API access","rank":7,"model_id":"zhipuai/glm-5.3","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3.toml"},{"model":"Kimi K3","score":0.6851,"effort":"max","percent":68.5,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":8,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"Grok 4.6","score":0.6748,"effort":"medium","percent":67.5,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"GPT-5.6 Luna","score":0.6719,"effort":"max","percent":67.2,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"GPT-5.5","score":0.6704,"effort":"xhigh","percent":67.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.7 Flash","score":0.6549,"effort":"medium","percent":65.5,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"GLM-5.3-Flash","score":0.6339,"effort":"max","percent":63.4,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":13,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Claude Opus 4.8","score":0.5897,"effort":null,"percent":59.0,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Qwen3.8 Max","score":0.5746,"effort":"xhigh","percent":57.5,"date":"2026-08-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":15,"model_id":"alibaba/qwen3.8-max","lab":"alibaba","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-max.toml"},{"model":"Muse Spark 1.2","score":0.5487,"effort":"xhigh","percent":54.9,"date":"2026-08-05","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":16,"model_id":"meta/muse-spark-1.2","lab":"meta","input_modalities":["text","image","video","pdf","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.2.toml"},{"model":"Claude Sonnet 5","score":0.5385,"effort":"max","percent":53.8,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":17,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Grok 4.5","score":0.5376,"effort":"high","percent":53.8,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Muse Spark 1.1","score":0.5332,"effort":null,"percent":53.3,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"GPT-5.4","score":0.5177,"effort":"xhigh","percent":51.8,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":20,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"Gemini 3.6 Flash","score":0.4668,"effort":null,"percent":46.7,"date":"2026-07-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21,"model_id":"google/gemini-3.6-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.6-flash.toml"},{"model":"GLM-5.2","score":0.4378,"effort":null,"percent":43.8,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":22,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Gemini 3.5 Flash","score":0.3739,"effort":null,"percent":37.4,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":23,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Kimi K2.7 Code","score":0.3053,"effort":null,"percent":30.5,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":24,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Claude Sonnet 4.6","score":0.2993,"effort":"high","percent":29.9,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":25,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.1173,"effort":null,"percent":11.7,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":26,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"}]},{"id":"scienceqa","name":"ScienceQA","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.25,"released":"2022-09-20","entrants":23,"leader":{"model":"Phi-3.5-vision-instruct","score":0.913,"effort":null,"percent":91.3,"date":"2024-08-16","organization":null,"countries":[],"accessibility":null,"rank":1},"rows":[{"model":"Phi-3.5-vision-instruct","score":0.913,"effort":null,"percent":91.3,"date":"2024-08-16","organization":null,"countries":[],"accessibility":null,"rank":1},{"model":"GPT-4o (May 2024)","score":0.885,"effort":null,"percent":88.5,"date":"2024-05-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Gemini 1.0 Pro Vision","score":0.797,"effort":null,"percent":79.7,"date":"2024-01-04","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"falcon-11B-vlm","score":0.749,"effort":null,"percent":74.9,"date":"2024-05-21","organization":null,"countries":[],"accessibility":null,"rank":4},{"model":"BLIP-2 (Q-Former)","score":0.7417,"effort":null,"percent":74.2,"date":"2023-02-06","organization":"Salesforce Research","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":5},{"model":"InstructGPT 175B","score":0.7404,"effort":null,"percent":74.0,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"llama3-llava-next-8b","score":0.737,"effort":null,"percent":73.7,"date":"2024-04-20","organization":null,"countries":[],"accessibility":null,"rank":7},{"model":"llava-v1.6-vicuna-13b","score":0.736,"effort":null,"percent":73.6,"date":"2024-01-31","organization":null,"countries":[],"accessibility":null,"rank":8},{"model":"llava-v1.6-mistral-7b","score":0.728,"effort":null,"percent":72.8,"date":"2024-01-31","organization":null,"countries":[],"accessibility":null,"rank":9},{"model":"MM1-7B-Chat","score":0.726,"effort":null,"percent":72.6,"date":"2024-03-14","organization":null,"countries":[],"accessibility":null,"rank":10},{"model":"Claude 3 Haiku","score":0.72,"effort":null,"percent":72.0,"date":"2024-03-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"llava-v1.6-vicuna-7b","score":0.706,"effort":null,"percent":70.6,"date":"2024-01-31","organization":null,"countries":[],"accessibility":null,"rank":12},{"model":"InternVL-Chat-ViT-6B-Vicuna-13B","score":0.701,"effort":null,"percent":70.1,"date":"2024-08-16","organization":null,"countries":[],"accessibility":null,"rank":13},{"model":"MM1-3B-Chat","score":0.694,"effort":null,"percent":69.4,"date":"2024-03-14","organization":null,"countries":[],"accessibility":null,"rank":14},{"model":"Qwen-VL-Chat","score":0.682,"effort":null,"percent":68.2,"date":"2023-08-20","organization":null,"countries":[],"accessibility":null,"rank":15},{"model":"llava-v1.5-7b","score":0.668,"effort":null,"percent":66.8,"date":"2023-10-05","organization":null,"countries":[],"accessibility":null,"rank":16},{"model":"InternVL-Chat-ViT-6B-Vicuna-7B","score":0.662,"effort":null,"percent":66.2,"date":"2023-12-25","organization":null,"countries":[],"accessibility":null,"rank":17},{"model":"instructblip-vicuna-13b","score":0.631,"effort":null,"percent":63.1,"date":"2023-12-25","organization":null,"countries":[],"accessibility":null,"rank":18},{"model":"instructblip-vicuna-7b","score":0.605,"effort":null,"percent":60.5,"date":"2023-05-22","organization":null,"countries":[],"accessibility":null,"rank":19},{"model":"Llama 2-13B","score":0.5578,"effort":null,"percent":55.8,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"LLaMA-13B","score":0.4333,"effort":null,"percent":43.3,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":21},{"model":"Llama 2-7B","score":0.4308,"effort":null,"percent":43.1,"date":"2023-07-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":22},{"model":"LLaMA-7B","score":0.3619,"effort":null,"percent":36.2,"date":"2023-02-24","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (non-commercial)","rank":23}]},{"id":"surface-evolver-bench","name":"Surface Evolver Bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-19","entrants":23,"leader":{"model":"Claude Fable 5","score":0.95,"effort":"high","percent":95.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.95,"effort":"high","percent":95.0,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Kimi K3","score":0.95,"effort":"unknown","percent":95.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":2,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GPT-5.6 Sol","score":0.9313,"effort":"xhigh","percent":93.1,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.5","score":0.8812,"effort":"high","percent":88.1,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Opus 4.8","score":0.875,"effort":null,"percent":87.5,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.6 Terra","score":0.8375,"effort":"xhigh","percent":83.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.6-terra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-terra.toml"},{"model":"Grok 4.5","score":0.7438,"effort":"high","percent":74.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"GPT-5.6 Luna","score":0.6188,"effort":"medium","percent":61.9,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.6-luna","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-luna.toml"},{"model":"Claude Sonnet 5","score":0.6,"effort":"medium","percent":60.0,"date":"2026-06-30","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"anthropic/claude-sonnet-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-5.toml"},{"model":"Gemini 3.5 Flash","score":0.5813,"effort":null,"percent":58.1,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"GLM-5.2","score":0.5563,"effort":null,"percent":55.6,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"MiniMax-M3","score":0.55,"effort":null,"percent":55.0,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":12},{"model":"Muse Spark 1.1","score":0.525,"effort":"high","percent":52.5,"date":"2026-07-09","organization":"Meta AI","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"meta/muse-spark-1.1","lab":"meta","input_modalities":["text","image","pdf","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/meta/muse-spark-1.1.toml"},{"model":"GLM-5.3-Flash","score":0.525,"effort":"max","percent":52.5,"date":"2026-08-20","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14,"model_id":"zhipuai/glm-5.3-flash","lab":"zhipuai","input_modalities":["text","image","video","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.3-flash.toml"},{"model":"Kimi K2.7 Code","score":0.4875,"effort":null,"percent":48.8,"date":"2026-06-12","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":15,"model_id":"moonshotai/kimi-k2.7-code","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.7-code.toml"},{"model":"Qwen3.8 27B","score":0.45,"effort":"xhigh","percent":45.0,"date":"2026-08-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16,"model_id":"alibaba/qwen3.8-27b","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.8-27b.toml"},{"model":"Qwen3.6 35B-A3B","score":0.4438,"effort":null,"percent":44.4,"date":"2026-04-14","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":17,"model_id":"alibaba/qwen3.6-35b-a3b","lab":"alibaba","input_modalities":["text","image","video","audio"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.6-35b-a3b.toml"},{"model":"DeepSeek v4 Pro","score":0.4,"effort":"high","percent":40.0,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Gemma 4 31B IT","score":0.3062,"effort":null,"percent":30.6,"date":"2026-04-02","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":19,"model_id":"google/gemma-4-31b-it","lab":"google","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemma-4-31b-it.toml"},{"model":"Mistral Medium 3.5","score":0.2687,"effort":null,"percent":26.9,"date":"2026-04-28","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (restricted use)","rank":20,"model_id":"mistral/mistral-medium-2604","lab":"mistral","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/mistral/mistral-medium-2604.toml"},{"model":"gpt-oss-120b","score":0.25,"effort":null,"percent":25.0,"date":"2025-08-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":21,"model_id":"openai/gpt-oss-120b","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-oss-120b.toml"},{"model":"Laguna M.1","score":0.1562,"effort":null,"percent":15.6,"date":"2026-04-28","organization":"Poolside","countries":["United States of America"],"accessibility":null,"rank":22},{"model":"Trinity Large Thinking","score":0.1562,"effort":null,"percent":15.6,"date":"2026-04-01","organization":"Arcee AI","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":23}]},{"id":"cybench","name":"Cybench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2024-08-15","entrants":22,"leader":{"model":"Claude Opus 4.6","score":0.93,"effort":"unknown thinking","percent":93.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},"rows":[{"model":"Claude Opus 4.6","score":0.93,"effort":"unknown thinking","percent":93.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Claude Opus 4.5","score":0.82,"effort":"unknown thinking","percent":82.0,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Claude Sonnet 4.5","score":0.6,"effort":"unknown thinking","percent":60.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.55,"effort":null,"percent":55.0,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Grok 4","score":0.43,"effort":null,"percent":43.0,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Claude Opus 4.1","score":0.42,"effort":null,"percent":42.0,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Grok 4.1","score":0.39,"effort":null,"percent":39.0,"date":"2025-11-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"Claude Opus 4","score":0.38,"effort":null,"percent":38.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-0.toml"},{"model":"Claude Sonnet 4","score":0.35,"effort":null,"percent":35.0,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Grok 4 Fast","score":0.3,"effort":null,"percent":30.0,"date":"2025-09-19","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"o3-mini","score":0.225,"effort":"medium","percent":22.5,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Claude 3.7 Sonnet","score":0.2,"effort":null,"percent":20.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"GPT-4.5 Preview (Feb 2025)","score":0.175,"effort":null,"percent":17.5,"date":"2025-02-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"Claude 3.5 Sonnet (Jun 2024)","score":0.175,"effort":null,"percent":17.5,"date":"2024-06-20","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14},{"model":"GPT-4o (Nov 2024)","score":0.125,"effort":null,"percent":12.5,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15},{"model":"Claude 3 Opus","score":0.1,"effort":null,"percent":10.0,"date":"2024-02-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":16},{"model":"o1-mini","score":0.1,"effort":"medium","percent":10.0,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":17},{"model":"o1-preview","score":0.1,"effort":null,"percent":10.0,"date":"2024-09-12","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":18},{"model":"Mixtral 8x22B","score":0.075,"effort":null,"percent":7.5,"date":"2024-04-17","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Llama 3.1-405B","score":0.075,"effort":null,"percent":7.5,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":20},{"model":"Gemini 1.5 Pro (Feb 2024)","score":0.075,"effort":null,"percent":7.5,"date":"2024-02-15","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":21},{"model":"Llama 3-70B","score":0.05,"effort":null,"percent":5.0,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":22}]},{"id":"ebr-bench","name":"EBR-bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-07-01","entrants":21,"leader":{"model":"GPT-6 Astra","score":0.7619,"effort":"max","percent":76.2,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},"rows":[{"model":"GPT-6 Astra","score":0.7619,"effort":"max","percent":76.2,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Fable 5.1","score":0.5714,"effort":"max","percent":57.1,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Opus 5","score":0.4571,"effort":"max","percent":45.7,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.6 Sol","score":0.4476,"effort":"max","percent":44.8,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Fable 5","score":0.3952,"effort":"max","percent":39.5,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.5","score":0.3429,"effort":"xhigh","percent":34.3,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Grok 4.6","score":0.3048,"effort":"xhigh","percent":30.5,"date":"2026-08-12","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"xai/grok-4.6","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.6.toml"},{"model":"Claude Opus 4.8","score":0.2857,"effort":null,"percent":28.6,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.4","score":0.254,"effort":"xhigh","percent":25.4,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GPT-5.2","score":0.2302,"effort":"xhigh","percent":23.0,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Opus 4.7","score":0.1905,"effort":"max","percent":19.0,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"Claude Opus 4.5 (128k thinking)","score":0.1429,"effort":null,"percent":14.3,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Gemini 3.1 Pro Preview","score":0.1429,"effort":null,"percent":14.3,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":13,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Opus 4.6","score":0.127,"effort":"max","percent":12.7,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"GPT-5","score":0.127,"effort":"high","percent":12.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"GLM-5.2","score":0.0952,"effort":null,"percent":9.5,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":16,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Qwen3.7 Max","score":0.0952,"effort":null,"percent":9.5,"date":"2026-05-19","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":17,"model_id":"alibaba/qwen3.7-max","lab":"alibaba","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-max.toml"},{"model":"Claude Opus 4.1","score":0.0794,"effort":null,"percent":7.9,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":18,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Gemini 3.5 Flash","score":0.0476,"effort":null,"percent":4.8,"date":"2026-05-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":19,"model_id":"google/gemini-3.5-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.5-flash.toml"},{"model":"Kimi K2.6","score":0.0238,"effort":null,"percent":2.4,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":20,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.0238,"effort":null,"percent":2.4,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":21}]},{"id":"cl-bench","name":"CL-bench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-02-03","entrants":21,"leader":{"model":"GPT-5.4","score":0.279,"effort":"xhigh","percent":27.9,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},"rows":[{"model":"GPT-5.4","score":0.279,"effort":"xhigh","percent":27.9,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GPT-5.1","score":0.237,"effort":"high","percent":23.7,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Grok 4.20","score":0.222,"effort":null,"percent":22.2,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Claude Opus 4.5","score":0.211,"effort":"unknown thinking","percent":21.1,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.208,"effort":null,"percent":20.8,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Opus 4.6","score":0.207,"effort":"unknown thinking","percent":20.7,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Qwen 3.6 Plus (2026-04-02)","score":0.203,"effort":null,"percent":20.3,"date":"2026-03-31","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":7},{"model":"Qwen3.5 Plus","score":0.198,"effort":null,"percent":19.8,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":8,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Kimi K2.5","score":0.193,"effort":null,"percent":19.3,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"GLM-5","score":0.187,"effort":null,"percent":18.7,"date":"2026-02-11","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":10,"model_id":"zhipuai/glm-5","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.toml"},{"model":"GPT-5.2","score":0.182,"effort":"unknown thinking","percent":18.2,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"o3","score":0.178,"effort":"high","percent":17.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Kimi K2 Thinking (Together)","score":0.176,"effort":null,"percent":17.6,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":13},{"model":"GLM-4.7 (Novita)","score":0.159,"effort":null,"percent":15.9,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14},{"model":"Gemini 3 Pro Preview","score":0.158,"effort":null,"percent":15.8,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":15,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"MiMo-V2-Pro","score":0.157,"effort":null,"percent":15.7,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"API access","rank":16},{"model":"Qwen3-Max-Instruct","score":0.145,"effort":null,"percent":14.5,"date":"2025-09-24","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":17},{"model":"DeepSeek-V3.2-Exp","score":0.132,"effort":null,"percent":13.2,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":18},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.124,"effort":null,"percent":12.4,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":19},{"model":"Kimi K2 Thinking","score":0.119,"effort":null,"percent":11.9,"date":"2025-11-06","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (restricted use)","rank":20,"model_id":"moonshotai/kimi-k2-thinking","lab":"moonshotai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2-thinking.toml"},{"model":"MiniMax-M2.5","score":0.114,"effort":null,"percent":11.4,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":21}]},{"id":"cadeval","name":"CadEval","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-04-22","entrants":15,"leader":{"model":"o3","score":0.74,"effort":"medium","percent":74.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},"rows":[{"model":"o3","score":0.74,"effort":"medium","percent":74.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Gemini 2.5 Pro Preview (Mar 2025)","score":0.64,"effort":null,"percent":64.0,"date":"2025-03-31","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"o4-mini","score":0.62,"effort":"medium","percent":62.0,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"o1","score":0.56,"effort":"medium","percent":56.0,"date":"2024-12-17","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"openai/o1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o1.toml"},{"model":"Claude 3.7 Sonnet","score":0.54,"effort":null,"percent":54.0,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"o3-mini","score":0.54,"effort":"medium","percent":54.0,"date":"2025-01-31","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/o3-mini","lab":"openai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3-mini.toml"},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.48,"effort":null,"percent":48.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"GPT-4.1","score":0.42,"effort":null,"percent":42.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-4.1","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1.toml"},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.34,"effort":null,"percent":34.0,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Claude 3.5 Haiku (Oct 2024)","score":0.32,"effort":null,"percent":32.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.3,"effort":null,"percent":30.0,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":11},{"model":"GPT-4o (Aug 2024)","score":0.26,"effort":null,"percent":26.0,"date":"2024-08-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Zoo Ml-Ephant","score":0.2,"effort":null,"percent":20.0,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":13},{"model":"GPT-4.1 mini","score":0.16,"effort":null,"percent":16.0,"date":"2025-04-14","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":14,"model_id":"openai/gpt-4.1-mini","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-4.1-mini.toml"},{"model":"Claude 3 Haiku","score":0.12,"effort":null,"percent":12.0,"date":"2024-03-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":15}]},{"id":"the-agent-company","name":"The Agent Company","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2024-12-18","entrants":14,"leader":{"model":"DeepSeek-V3.2-Exp","score":0.429,"effort":null,"percent":42.9,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":1},"rows":[{"model":"DeepSeek-V3.2-Exp","score":0.429,"effort":null,"percent":42.9,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":1},{"model":"Gemini 2.5 Flash (Sep 2025)","score":0.411,"effort":null,"percent":41.1,"date":"2025-09-25","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Claude Sonnet 4","score":0.331,"effort":null,"percent":33.1,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.7 Sonnet","score":0.309,"effort":null,"percent":30.9,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.303,"effort":null,"percent":30.3,"date":"2025-05-06","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Claude 3.5 Sonnet (Oct 2024)","score":0.24,"effort":null,"percent":24.0,"date":"2024-10-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"Gemini 2.0 Flash (Feb 2025)","score":0.114,"effort":null,"percent":11.4,"date":"2025-02-05","organization":"Google DeepMind,Google","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"GPT-4o (Nov 2024)","score":0.086,"effort":null,"percent":8.6,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8},{"model":"Llama 3.1-405B","score":0.074,"effort":null,"percent":7.4,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":9},{"model":"Llama 3.1-70B","score":0.069,"effort":null,"percent":6.9,"date":"2024-07-23","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":10},{"model":"Qwen2.5-72B","score":0.057,"effort":null,"percent":5.7,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"Gemini 1.5 Pro (Sept 2024)","score":0.034,"effort":null,"percent":3.4,"date":"2024-09-24","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":12},{"model":"Amazon Nova Pro","score":0.017,"effort":null,"percent":1.7,"date":"2024-12-03","organization":"Amazon","countries":["United States of America"],"accessibility":"API access","rank":13},{"model":"Qwen2-72B","score":0.011,"effort":null,"percent":1.1,"date":"2024-06-07","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14}]},{"id":"cl-bench-life","name":"CL-bench Life","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-04-29","entrants":14,"leader":{"model":"GPT-5.5","score":0.222,"effort":"high","percent":22.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},"rows":[{"model":"GPT-5.5","score":0.222,"effort":"high","percent":22.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"GPT-5.4","score":0.217,"effort":"xhigh","percent":21.7,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GPT-5.1","score":0.173,"effort":"high","percent":17.3,"date":"2025-11-13","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.1","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.1.toml"},{"model":"Claude Opus 4.6","score":0.17,"effort":"high","percent":17.0,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.169,"effort":null,"percent":16.9,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"DeepSeek v4 Pro","score":0.135,"effort":"high","percent":13.5,"date":"2026-04-24","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6,"model_id":"deepseek/deepseek-v4-pro","lab":"deepseek","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/deepseek/deepseek-v4-pro.toml"},{"model":"Kimi K2.5","score":0.132,"effort":null,"percent":13.2,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":7},{"model":"Qwen3.5 Plus","score":0.124,"effort":null,"percent":12.4,"date":"2026-02-16","organization":"Alibaba","countries":["China"],"accessibility":null,"rank":8,"model_id":"alibaba/qwen3.5-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.5-plus.toml"},{"model":"Grok 4.20","score":0.119,"effort":null,"percent":11.9,"date":"2026-02-17","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"GLM-4.7 (Novita)","score":0.109,"effort":null,"percent":10.9,"date":"2025-12-22","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":10},{"model":"DeepSeek-V3.2-Exp","score":0.095,"effort":null,"percent":9.5,"date":"2025-09-29","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":11},{"model":"DeepSeek-V3.2 (Thinking; Novita)","score":0.074,"effort":null,"percent":7.4,"date":"2025-12-01","organization":"DeepSeek","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":12},{"model":"MiMo-V2-Pro","score":0.069,"effort":null,"percent":6.9,"date":null,"organization":"Xiaomi Corp","countries":["China"],"accessibility":"API access","rank":13},{"model":"MiniMax-M2.5","score":0.063,"effort":null,"percent":6.3,"date":"2026-02-12","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":14}]},{"id":"anli","name":"ANLI","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.3333333333333333,"released":"2019-10-31","entrants":11,"leader":{"model":"phi-3-small 7.4B","score":0.581,"effort":null,"percent":58.1,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},"rows":[{"model":"phi-3-small 7.4B","score":0.581,"effort":null,"percent":58.1,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},{"model":"GPT-3.5 Turbo (Nov 2023)","score":0.581,"effort":null,"percent":58.1,"date":"2023-11-06","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Llama 3-8B","score":0.573,"effort":null,"percent":57.3,"date":"2024-04-18","organization":"Meta AI","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":3},{"model":"phi-3-medium 14B","score":0.558,"effort":null,"percent":55.8,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":4},{"model":"Mixtral 8x7B","score":0.552,"effort":null,"percent":55.2,"date":"2023-12-11","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":5},{"model":"phi-3-mini 3.8B","score":0.528,"effort":null,"percent":52.8,"date":"2024-04-23","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":6},{"model":"Gemma 7B","score":0.487,"effort":null,"percent":48.7,"date":"2024-02-21","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"Open weights (restricted use)","rank":7},{"model":"Mistral 7B v0.1","score":0.471,"effort":null,"percent":47.1,"date":"2023-09-27","organization":"Mistral AI","countries":["France"],"accessibility":"Open weights (unrestricted)","rank":8},{"model":"Phi-2","score":0.425,"effort":null,"percent":42.5,"date":"2023-12-12","organization":"Microsoft","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":9},{"model":"Megatron-Turing NLG 530B","score":0.397,"effort":null,"percent":39.7,"date":"2021-10-11","organization":"Microsoft,NVIDIA","countries":["United States of America"],"accessibility":"Unreleased","rank":10},{"model":"InstructGPT 175B","score":0.354,"effort":null,"percent":35.4,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11}]},{"id":"posttrainbench","name":"PostTrainBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-03-11","entrants":11,"leader":{"model":"Claude Fable 5","score":0.4179,"effort":"max","percent":41.8,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.4179,"effort":"max","percent":41.8,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-5.6 Sol","score":0.3623,"effort":"max","percent":36.2,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 5","score":0.3504,"effort":"unknown","percent":35.0,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"Claude Opus 4.8","score":0.3384,"effort":null,"percent":33.8,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Kimi K3","score":0.3196,"effort":"unknown","percent":32.0,"date":"2026-07-16","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":5,"model_id":"moonshotai/kimi-k3","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k3.toml"},{"model":"GLM-5.2","score":0.317,"effort":null,"percent":31.7,"date":"2026-06-16","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6,"model_id":"zhipuai/glm-5.2","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.2.toml"},{"model":"Claude Opus 4.7","score":0.2856,"effort":"xhigh","percent":28.6,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.5","score":0.2723,"effort":"xhigh","percent":27.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Grok 4.5","score":0.2345,"effort":"high","percent":23.4,"date":"2026-07-08","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"xai/grok-4.5","lab":"xai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/xai/grok-4.5.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.2199,"effort":null,"percent":22.0,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"GPT-5.4","score":0.19,"effort":"high","percent":19.0,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"}]},{"id":"remote-labor-index","name":"Remote Labor Index","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-10-29","entrants":11,"leader":{"model":"Claude Fable 5","score":0.161,"effort":null,"percent":16.1,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},"rows":[{"model":"Claude Fable 5","score":0.161,"effort":null,"percent":16.1,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"Claude Opus 4.8","score":0.0833,"effort":null,"percent":8.3,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"GPT-5.5","score":0.0625,"effort":"unknown thinking","percent":6.2,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.7 Flash","score":0.05,"effort":"unknown","percent":5.0,"date":"2026-08-13","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"google/gemini-3.7-flash","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.7-flash.toml"},{"model":"Claude Opus 4.6","score":0.0417,"effort":"unknown thinking","percent":4.2,"date":"2026-02-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-opus-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-6.toml"},{"model":"Claude Opus 4.5","score":0.0375,"effort":"unknown thinking","percent":3.8,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-opus-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-5.toml"},{"model":"GPT-5.2","score":0.025,"effort":"medium","percent":2.5,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Sonnet 4.5","score":0.0208,"effort":"unknown thinking","percent":2.1,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-sonnet-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-5.toml"},{"model":"GPT-5","score":0.0167,"effort":"unknown thinking","percent":1.7,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":9,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"Gemini 3 Pro Preview","score":0.0125,"effort":null,"percent":1.2,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":10,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"Gemini 2.5 Pro Preview (Jun 2025)","score":0.0083,"effort":null,"percent":0.8,"date":"2025-06-05","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":11}]},{"id":"gdpval","name":"GDPval","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2025-09-25","entrants":11,"leader":{"model":"GPT-5.2","score":0.497,"effort":"none","percent":49.7,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},"rows":[{"model":"GPT-5.2","score":0.497,"effort":"none","percent":49.7,"date":"2025-12-11","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"openai/gpt-5.2","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.2.toml"},{"model":"Claude Opus 4.5 (no thinking)","score":0.455,"effort":null,"percent":45.5,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Claude Opus 4.1","score":0.436,"effort":null,"percent":43.6,"date":"2025-08-05","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-4-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-1.toml"},{"model":"Claude Sonnet 4.5 (no thinking)","score":0.425,"effort":null,"percent":42.5,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Gemini 3 Pro Preview","score":0.403,"effort":null,"percent":40.3,"date":"2025-11-18","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"google/gemini-3-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3-pro-preview.toml"},{"model":"GPT-5","score":0.348,"effort":"medium","percent":34.8,"date":"2025-08-07","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.toml"},{"model":"o3","score":0.308,"effort":"medium","percent":30.8,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"o4-mini","score":0.253,"effort":"high","percent":25.3,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/o4-mini","lab":"openai","input_modalities":["text","image"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o4-mini.toml"},{"model":"Gemini 2.5 Pro (Jun 2025)","score":0.233,"effort":null,"percent":23.3,"date":"2025-06-17","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":9},{"model":"Grok 4","score":0.211,"effort":null,"percent":21.1,"date":"2025-07-09","organization":"xAI","countries":["United States of America"],"accessibility":"API access","rank":10},{"model":"GPT-4o (Nov 2024)","score":0.099,"effort":null,"percent":9.9,"date":"2024-11-20","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":11}]},{"id":"osworld","name":"OSWorld","in_index":true,"scale":0.01,"ceiling":1.0,"random_baseline":0.0,"released":"2024-04-11","entrants":9,"leader":{"model":"Claude Sonnet 4.6 (no thinking)","score":72.1,"effort":null,"percent":null,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1},"rows":[{"model":"Claude Sonnet 4.6 (no thinking)","score":72.1,"effort":null,"percent":null,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1},{"model":"Claude Opus 4.5 (no thinking)","score":66.3,"effort":null,"percent":null,"date":"2025-11-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2},{"model":"Kimi K2.5","score":63.3,"effort":null,"percent":null,"date":"2026-01-27","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":3},{"model":"Claude Sonnet 4.5 (no thinking)","score":62.9,"effort":null,"percent":null,"date":"2025-09-29","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4},{"model":"Claude Sonnet 4","score":43.9,"effort":null,"percent":null,"date":"2025-05-22","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"anthropic/claude-sonnet-4-0","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-0.toml"},{"model":"Claude 3.7 Sonnet","score":35.8,"effort":null,"percent":null,"date":"2025-02-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6},{"model":"computer-use-preview-2025-03-11","score":31.3,"effort":null,"percent":null,"date":"2025-03-11","organization":null,"countries":[],"accessibility":null,"rank":7},{"model":"o3","score":23.0,"effort":"medium","percent":null,"date":"2025-04-16","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"openai/o3","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/o3.toml"},{"model":"Qwen2.5-72B","score":5.0,"effort":null,"percent":null,"date":"2024-09-19","organization":"Alibaba","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":9}]},{"id":"exploitbench","name":"ExploitBench","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-05-13","entrants":9,"leader":{"model":"Claude Mythos Preview (Early)","score":0.738,"effort":null,"percent":73.8,"date":"2026-04-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"Unreleased","rank":1},"rows":[{"model":"Claude Mythos Preview (Early)","score":0.738,"effort":null,"percent":73.8,"date":"2026-04-07","organization":"Anthropic","countries":["United States of America"],"accessibility":"Unreleased","rank":1},{"model":"GPT-5.5","score":0.474,"effort":"unknown thinking","percent":47.4,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Opus 4.7 (no thinking)","score":0.265,"effort":null,"percent":26.5,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3},{"model":"Gemini 3.1 Pro Preview","score":0.261,"effort":null,"percent":26.1,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"},{"model":"Claude Sonnet 4.6 (no thinking)","score":0.236,"effort":null,"percent":23.6,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":5},{"model":"Kimi K2.6","score":0.184,"effort":null,"percent":18.4,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":6,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"GLM-5.1","score":0.181,"effort":null,"percent":18.1,"date":"2026-04-07","organization":"Z.ai (Zhipu AI)","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":7,"model_id":"zhipuai/glm-5.1","lab":"zhipuai","input_modalities":["text"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/zhipuai/glm-5.1.toml"},{"model":"Claude Haiku 4.5","score":0.137,"effort":"unknown thinking","percent":13.7,"date":"2025-10-15","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"anthropic/claude-haiku-4-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-haiku-4-5.toml"},{"model":"MiniMax-M2.7","score":0.133,"effort":null,"percent":13.3,"date":"2026-03-18","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (non-commercial)","rank":9}]},{"id":"osworld-2-0","name":"OSWorld 2.0","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-26","entrants":9,"leader":{"model":"Claude Opus 5","score":0.3143,"effort":"max","percent":31.4,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},"rows":[{"model":"Claude Opus 5","score":0.3143,"effort":"max","percent":31.4,"date":"2026-07-24","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-opus-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-5.toml"},{"model":"GPT-5.6 Sol","score":0.2734,"effort":"max","percent":27.3,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"Claude Opus 4.8","score":0.206,"effort":null,"percent":20.6,"date":"2026-05-28","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"anthropic/claude-opus-4-8","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-8.toml"},{"model":"Claude Opus 4.7","score":0.182,"effort":"max","percent":18.2,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.5","score":0.13,"effort":"xhigh","percent":13.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Claude Sonnet 4.6","score":0.093,"effort":"medium","percent":9.3,"date":"2026-02-17","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"anthropic/claude-sonnet-4-6","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-sonnet-4-6.toml"},{"model":"MiniMax-M3","score":0.046,"effort":null,"percent":4.6,"date":"2026-06-01","organization":"MiniMax","countries":["China"],"accessibility":"Open weights (restricted use)","rank":7},{"model":"Kimi K2.6","score":0.046,"effort":null,"percent":4.6,"date":"2026-04-20","organization":"Moonshot","countries":["China"],"accessibility":"Open weights (unrestricted)","rank":8,"model_id":"moonshotai/kimi-k2.6","lab":"moonshotai","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/moonshotai/kimi-k2.6.toml"},{"model":"Qwen3.7 Plus","score":0.028,"effort":null,"percent":2.8,"date":"2026-06-02","organization":"Alibaba","countries":["China"],"accessibility":"API access","rank":9,"model_id":"alibaba/qwen3.7-plus","lab":"alibaba","input_modalities":["text","image","video"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/alibaba/qwen3.7-plus.toml"}]},{"id":"mirrorcode","name":"MirrorCode","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2026-06-26","entrants":8,"leader":{"model":"Claude Fable 5.1","score":0.733,"effort":"high","percent":73.3,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},"rows":[{"model":"Claude Fable 5.1","score":0.733,"effort":"high","percent":73.3,"date":"2026-09-01","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":1,"model_id":"anthropic/claude-fable-5-1","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5-1.toml"},{"model":"Claude Fable 5","score":0.6389,"effort":"high","percent":63.9,"date":"2026-06-09","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":2,"model_id":"anthropic/claude-fable-5","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-fable-5.toml"},{"model":"GPT-6 Astra","score":0.467,"effort":"high","percent":46.7,"date":"2026-09-03","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":3,"model_id":"openai/gpt-6-astra","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-6-astra.toml"},{"model":"Claude Opus 4.7","score":0.3111,"effort":"high","percent":31.1,"date":"2026-04-16","organization":"Anthropic","countries":["United States of America"],"accessibility":"API access","rank":4,"model_id":"anthropic/claude-opus-4-7","lab":"anthropic","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/anthropic/claude-opus-4-7.toml"},{"model":"GPT-5.6 Sol","score":0.2,"effort":"high","percent":20.0,"date":"2026-07-09","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":5,"model_id":"openai/gpt-5.6-sol","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.6-sol.toml"},{"model":"GPT-5.4","score":0.1556,"effort":"high","percent":15.6,"date":"2026-03-05","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":6,"model_id":"openai/gpt-5.4","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.4.toml"},{"model":"GPT-5.5","score":0.1,"effort":"high","percent":10.0,"date":"2026-04-23","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7,"model_id":"openai/gpt-5.5","lab":"openai","input_modalities":["text","image","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/openai/gpt-5.5.toml"},{"model":"Gemini 3.1 Pro Preview","score":0.0889,"effort":null,"percent":8.9,"date":"2026-02-19","organization":"Google DeepMind","countries":["United States of America"],"accessibility":"API access","rank":8,"model_id":"google/gemini-3.1-pro-preview","lab":"google","input_modalities":["text","image","video","audio","pdf"],"output_modalities":["text"],"modality_source":"https://github.com/anomalyco/models.dev/blob/dev/models/google/gemini-3.1-pro-preview.toml"}]},{"id":"superglue","name":"SuperGLUE","in_index":true,"scale":1.0,"ceiling":1.0,"random_baseline":0.0,"released":"2019-05-02","entrants":8,"leader":{"model":"T5-11B","score":0.889,"effort":null,"percent":88.9,"date":null,"organization":"Google","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},"rows":[{"model":"T5-11B","score":0.889,"effort":null,"percent":88.9,"date":null,"organization":"Google","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":1},{"model":"T5-3B","score":0.864,"effort":null,"percent":86.4,"date":null,"organization":"Google","countries":["United States of America"],"accessibility":"Open weights (unrestricted)","rank":2},{"model":"Switch-Large","score":0.847,"effort":null,"percent":84.7,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":3},{"model":"T5-Large","score":0.827,"effort":null,"percent":82.7,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":4},{"model":"T5-Base","score":0.762,"effort":null,"percent":76.2,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":5},{"model":"Switch-Base","score":0.733,"effort":null,"percent":73.3,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":6},{"model":"InstructGPT 175B","score":0.718,"effort":null,"percent":71.8,"date":"2022-01-27","organization":"OpenAI","countries":["United States of America"],"accessibility":"API access","rank":7},{"model":"T5-Small","score":0.633,"effort":null,"percent":63.3,"date":null,"organization":null,"countries":[],"accessibility":null,"rank":8}]}]}