Deepseek V 3.1
DeepSeek-V3.1 is post-trained on the top of DeepSeek-V3.1-Base, which is built upon the original V3 base checkpoint through a two-phase long context extension approach, following the methodology outlined in the original DeepSeek-V3 report. We have expanded our dataset by collecting additional long documents and substantially extending both training phases.
Input
Output
Context
131K
Max Output
128K
Parameters
671B
Technical Specifications
Model TypeChat
Context Window131K
Max Output Tokens128K
Parameters671B
Training Cutoff—
Licensemit
Capabilities
Input Modalities
text
Output Modalities
text
Features
code_generationfunction_callingjson_modemathreasoningstreamingtext_generation
Benchmarks
6.3%
HLE
73.5%
GPQA
36.7%
SciCode
83.3%
MMLU Pro
49.7
Math
28.4
Coding
57.7%
LiveCodeBench
28.1
Intelligence
Resources & Links
Related Tools
Data sourced from official provider APIs and documentation
Last updated: May 5, 2026
Start building with the right model.
From model selection to production, one platform, no fragmentation.