Qwen3.5-397B-A17B: hardware requirements
Qwen3.5-397B-A17B needs about 244 GB of memory at Q4 quantization — a 241 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.
This is a mixture-of-experts model. Only a fraction of its 403B parameters are used for each token, so it runs faster than its size suggests — but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →
Will it run on your machine?
Download sizes by quantization
Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?
| Quant | Download | RAM @ 4K | RAM @ 32K | Source |
|---|---|---|---|---|
| Q4_K_Mrecommended | 241 GB | 244 GB | 248 GB | measured |
| Q5_K_M | 286 GB | 290 GB | 293 GB | estimated |
| Q6_K | 326 GB | 329 GB | 333 GB | measured |
| Q8_0 | 422 GB | 425 GB | 429 GB | measured |
| BF16 | 807 GB | 811 GB | 814 GB | estimated |
“Measured” sizes are read from published GGUF files (timteh673/Qwen3.5-397B-A17B-Opus-4.6-Reasoning-Uncensored-GGUF). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.