DeepSeek V4 Pro 0423
DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes
BENCHMARKS
Not in the Epoch Capabilities Index. Epoch benchmarks what it can run, so this is a gap in coverage, not a poor result.
RELEASE DATEApr 23, 2026
WEIGHTSOpen weights
CONTEXT WINDOW1M
MAX OUTPUT384K
REASONINGYes
TOOL CALLINGYes
INPUTStext
OUTPUTStext
Specifications from Models.dev. Availability and limits may differ by provider. Benchmark scores come from Epoch AI under CC BY 4.0.