What GPU do you need to run Qwen3.5 397B A17B?
397B params 256K max context VisionTool useReasoning
The smallest GPU memory that runs Qwen3.5 397B A17B well is 256 GB (Q4_K_M, e.g. M3 Ultra). 2 of 132 GPUs run it well. It's a mixture-of-experts model, so all weights must fit in memory but speed depends only on the active experts.
MMLU 87 SWE-bench Verified 76.2
VRAM needed per quantization
| Quantization | Weights | 16K context | 32K context | Full 256K |
|---|---|---|---|---|
| Q4_K_M | 227.3 GB | 227.8 GB | 228.3 GB | 234.8 GB |
| Q8_0 | 392.6 GB | 393.0 GB | 393.5 GB | 400.0 GB |
Weights plus fp16 KV cache. Leave ~1 GB headroom for the inference engine.
GPUs that can run Qwen3.5 397B A17B
Best quantization that fits on each GPU, with the context and speed it gets. For chips with several memory sizes, the smallest that works is shown.