Kimi K3

Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work

ECI STANDING#11 of 266
SCORE157.63155.38–160.45 CI
FRONTIER LAGLaunched 4.4 months behindScores about what GPT-5.4 Pro did in March 2026.
Full Epoch Capabilities Index ranking

From the lab

Related articleKimi K3 Tech Blog: Open Frontier Intelligence
RELEASE DATEJul 16, 2026
WEIGHTSOpen weights
CONTEXT WINDOW1.05M
MAX OUTPUT131.1K
REASONINGYes
TOOL CALLINGYes
INPUTStext, image, video
OUTPUTStext

Specifications from Models.dev. Availability and limits may differ by provider. Benchmark scores come from Epoch AI under CC BY 4.0.