Only with CPU offload
Qwen3.5 397B-A17B at Q4_K_M needs 227.4 GB but only 10.6 GB is addressable, so about 96% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 1.5 tokens per second — usable for batch work, painful for chat.
Flagship-class at 17B active. A 256 GB Mac Studio or a multi-GPU box at Q4.
Hardware requirements What hardware do I need to run Qwen3.5 397B-A17B? →At Q4_K_M with an 8K context: a 512 GB Mac (M3 Ultra). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.
Fits instead: Qwen3.5 9B (6.3 GB) · Qwen3.5 4B (3.5 GB)
The VRAM budget
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 398.8 GB | 399.6 GB | — | ~0.8 | −0.1% ppl | 389.0 GB over |
| Q6_K | 307.8 GB | 308.6 GB | — | ~1.1 | −0.4% ppl | 298.0 GB over |
| Q5_K_M | 266.0 GB | 266.8 GB | — | ~1.3 | −0.8% ppl | 256.2 GB over |
| Q4_K_M | 226.6 GB | 227.4 GB | — | ~1.5 | −1.9% ppl | 216.8 GB over |
| Q3_K_M | 183.4 GB | 184.3 GB | — | ~1.9 | −5.4% ppl | 173.7 GB over |
| Q2_K | 157.2 GB | 158.0 GB | — | ~2.2 | −15% ppl | 147.4 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 15 of its 60 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state.
How to run it
$ llama-server \
-hf Qwen/Qwen3.5-397B-A17B:Q4_K_M \
-c 8192 -ngl 2
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
Related models on a GeForce RTX 3060 12 GB
| Model | Total | Tok/s est. | |
|---|---|---|---|
| Qwen3.5 122B-A10B 125B MoE · same family |
Q4_K_M · 71.1 GB | ~2.8 | CPU offload |
| Qwen3.5 9B 9.65B · same family |
Q4_K_M · 6.3 GB | 40 | Runs great |
| Qwen3.5 4B 4.66B · same family |
Q4_K_M · 3.5 GB | 83 | Runs great |
| Qwen3.5 2B 2.27B · same family |
Q4_K_M · 2.0 GB | 171 | Runs great |
| Ornith 1.5 397B-A17B 397B MoE · similar size |
Q4_K_M · 224.1 GB | ~1.5 | CPU offload |
| GLM-5.3-Flash 320B-A18B 321B MoE · similar size |
Q4_K_M · 181.2 GB | ~1.4 | CPU offload |
| DeepSeek V4.1 Flash 552B-A16B 552B MoE · similar size |
Q4_K_M · 311.0 GB | — | Won't fit |
Qwen3.5 397B-A17B on other hardware
| Device | Total | Tok/s est. | |
|---|---|---|---|
| GeForce RTX 4060 Laptop 7.0 GB usable · 256 GB/s |
227.4 GB | — | Won't fit |
| M4 · 16 GB 10.7 GB usable · 120 GB/s |
227.4 GB | — | Won't fit |
| GeForce RTX 4090 22.4 GB usable · 1008 GB/s |
227.4 GB | ~1.6 | CPU offload |
| Radeon RX 7900 XTX 22.4 GB usable · 960 GB/s |
227.4 GB | ~1.6 | CPU offload |
| M3 Max · 64 GB 48.0 GB usable · 400 GB/s |
227.4 GB | — | Won't fit |
| M3 Ultra · 512 GB 384.0 GB usable · 800 GB/s |
227.4 GB | 22 | Runs great |