QwenQwen3.5-397B (MoE) on Apple M4 Ultra (192GB)
Yes — M4 Ultra (192GB) handles Qwen3.5-397B (MoE) well at Q3_K_M — 44 tok/s. Solid daily-driver performance on 192 GB VRAM.
Model Size
397B
Device VRAM
192 GB
Bandwidth
819 GB/s
Quantization
Q3_K_M
Performance by Quantization
OwnRig currently has one published compatibility entry for Qwen3.5-397B (MoE) on Apple M4 Ultra (192GB) at Q3_K_M. This is the best supported pairing we can stand behind today.
| Quantization | Speed | TTFT | Fits in VRAM | Rating | Confidence |
|---|---|---|---|---|---|
| Q3_K_M | 44 tok/s | 182ms | ✓ Yes | Good | estimated |
Notes
Q3_K_M
Q3_K_M fits in 192GB unified memory with limited headroom. Frontier MoE model: 397B total params, ~30B active per token. Consumer hardware requires heavy offloading.
About Qwen3.5-397B (MoE)
Qwen3.5-397B (MoE) (397B) is a chat, coding, ai coding, reasoning, multi-purpose model. Frontier-scale MoE with 397B total parameters and roughly 30B active per token. Requires an enormous memory footprint even when quantized, so local runs are limited to GB300-class hardware or aggressive offloading setups.
View all Qwen3.5-397B (MoE) hardware options →About Apple M4 Ultra (192GB)
Apple M4 Ultra (192GB) has 192 GB at 819 GB/s. Available in Mac Studio, Mac Pro.
See all models Apple M4 Ultra (192GB) can run →Estimate method: Estimated from frontier MoE reports and bandwidth scaling for ~30B-active models. Reference hardware source: github.com (2026-03-24)
Performance varies by driver version, inference engine, quantization method, context length, and system configuration. Figures shown are estimates based on community benchmarks and may not reflect your exact setup. Product names are trademarks of their respective owners. OwnRig is independent and not affiliated with any hardware or AI model provider.