models · qwen
Code-specialized 32B model with open-source SOTA HumanEval / code-completion quality.
| Weight file SHA-256 | d3bbdd1e787d7e36d63498f035d549e7515af8d6dce0385264a4c98509e0d76a |
|---|---|
| Method | hf-published-sha256 |
| Publisher | qwen |
|---|---|
| Origin | Recorded as a covered nation under 10 U.S.C. § 4872(f) — the PRC, Russia, Iran or North Korea. The registry did not record which. |
| Licence | Apache-2.0 |
| Parameters | 32,763,876,352 (32.8B) |
| Context window | 128K |
| Modality | text->text |
| Architecture | qwen2 |
| Layers | 64 |
| Hidden size | 5120 |
| Attention | GQA (40 heads, 8 KV) |
| Native precision | BF16 |
| Pulls recorded here | 480,013 |
| Downloads reported upstream | 1,484,515 |
| MMLU-Pro | 63.5 |
|---|---|
| GPQA Diamond | 41.7 |
| MATH-500 | 76.7 |
| LiveCodeBench | 29.5 |
| Humanity's Last Exam | 3.5 |
Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.
| Hardware | Build measured | Decode @ 8K | Memory @ 8K | Decode mode |
|---|---|---|---|---|
| NVIDIA DGX Spark 128GB | — | 3.5 tok/s | 67.8 GB | not stated |
Measured by local.ai, not by us: 1 run. Each names its engine pinned by image digest and the exact serve command, in the full record.
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.