models · deepseek-ai
Merge of DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct. Balanced coding and general.
| Weight file SHA-256 | cfdac3d5826bc0e99f8db4f9c5f325795527b33697e988381173193347bc54ed |
|---|---|
| Method | hf-published-sha256 |
| Publisher | deepseek-ai |
|---|---|
| Origin | China |
| Licence | DeepSeek-TOS |
| Publisher’s own licence statement | other |
| Parameters | 235,741,434,880 (235.7B) |
| Architecture | deepseek_v2 |
| Layers | 60 |
| Hidden size | 5120 |
| Attention | MLA (128 heads, 128 KV) |
| Native precision | BF16 |
| Pulls recorded here | 48,016 |
| Downloads reported upstream | 5,486 |
| MMLU-Pro | 66.6 |
|---|---|
| GPQA Diamond | 42.3 |
| MATH-500 | 76.3 |
Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.
Registry record as of 2026-06-11. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.