Question 1

Can I run Qwen3.6 35B A3B on a Apple M3 Max?

Accepted Answer

Yes — Apple M3 Max runs Qwen3.6 35B A3B fully on GPU at the FP16 quantization. Full GPU inference — 128 GB VRAM meets the 91 GB recommendation.

Question 2

What quantization of Qwen3.6 35B A3B should I use on a Apple M3 Max?

Accepted Answer

For 128 GB VRAM on the Apple M3 Max, the FP16 variant is the best fit. Estimated ~5.7 tokens/sec on the FP16 quantization.

Question 3

How fast does Qwen3.6 35B A3B run on Apple M3 Max?

Accepted Answer

Roughly 5.7 tokens/sec for FP16. Real speed depends on context length, backend (Ollama, llama.cpp, LM Studio), and KV cache size.

Question 4

What if Apple M3 Max is not enough for Qwen3.6 35B A3B?

Accepted Answer

Try a smaller quantization variant or run on CPU + RAM with the trade-off of slower generation.

Quantization	File Size	Min VRAM	Rec VRAM	Context	Verdict	Estimated tok/s
Q4_K_M	17.5 GB	20.1 GB	22.8 GB	8K / 8K	Full GPU	~18.3
Q5_K_M	21.9 GB	25.2 GB	28.5 GB	8K / 8K	Full GPU	~15.9
Q8_0	35 GB	40.3 GB	45.5 GB	8K / 8K	Full GPU	~10.9
FP16Best fit	70 GB	80.5 GB	91 GB	8K / 8K	Full GPU	~5.7

Can I Run Qwen3.6 35B A3B on Apple M3 Max?

Share this matchup