DeepSeek-V3.1
DeepSeek·
LLMsopen-weightFrontier
Overview
Incremental update with better instruction following and improved benchmark scores across the board.
Capabilities and innovations
671B total / 37B active parameters (MoE)Improved instruction following128K context windowEnhanced coding and reasoningRefined RLHF alignmentImproved multi-turn consistencyBetter long-context utilization
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 81.4% | unsourced |
| Humanity’s Last Exam (no tools) | 25.1% | unsourced |
| SWE-bench Verified | 54.7% | unsourced |
| MMLU | 87.6% | unsourced |
Architecture and hardware
- Parameters
- 671B total, 37B active per token (MoE)
- Estimated VRAM at Q4
- ~386 GB, Frontier class
- Quantization formats
- GGUF, GPTQ, FP8
- Recommended runtime
- vLLM
- License
- MIT
Links
More from DeepSeek
Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.