community pool / model answer
Qwen3.6-35B-A3B-NVFP4
chat · nvidia/Qwen3.6-35B-A3B-NVFP4
233.4 tok/s · model median · 19 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| GB10 Grace Blackwell 128GB | 4-bit | 106 tok/s | measured | 135 | - GB | 262,144 | vllm |
| RTX 3090 24GB ×2 | 4-bit | 233.4 tok/s | measured | 45 | 41.4 GB | 262,144 | vllm |
| RTX 3090 24GB ×3 | 4-bit | 231.3 tok/s | reported | - | - GB | 2,048 | vllm |
| RTX PRO 6000 Blackwell 96GB | 4-bit | 330.3 tok/s | measured | 5,579 | - GB | 262,144 | vllm |
| RTX PRO 6000 Blackwell 96GB ×4 | 4-bit | 325.9 tok/s | reported | 39 | - GB | 2,048 | vllm |
Context tested
The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.