Alibaba · 36B (3B active) · 專家混合（MoE）
Big-model quality at 3B-active speed — the mid-hardware sweet spot 看看你的 GPU 或 Mac 跑不跑得動 Qwen 3.6 35B-A3B——最低 20.1 GB，建議 33.5 GB。
2026-04256K context
專家混合（MoE）
專家總數： 256
啟用專家： 8
啟用參數： 3.0B
| 量化 | 位元 | VRAM | 品質 | 狀態 |
|---|---|---|---|---|
| Q2_K | 2 | 12 GB | low | — |
| Q3_K_M | 3 | 16.6 GB | moderate | — |
| Q4_K_M | 4 | 18.9 GB | good | — |
| Q5_K_M | 5 | 23.6 GB | good | — |
| Q6_K | 6 | 28.2 GB | excellent | — |
| Q8_0 | 8 | 37.4 GB | excellent | — |
| F16 | 16 | 74.3 GB | lossless | — |
Can I run Qwen 3.6 35B-A3B locally?
- Can I run Qwen 3.6 35B-A3B locally?
- Qwen 3.6 35B-A3B needs about 20.1 GB of memory at a minimum and 33.5 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does Qwen 3.6 35B-A3B need?
- At Q4_K_M, Qwen 3.6 35B-A3B uses about 18.9 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.