Compatibility check
Can I run Qwen3 ASR 1.7B HF on Apple M4 Max (40-core GPU)?
Yes — it runs.best quant:
Decode speed
315 tok/sest
Usable context
262K
Memory
128 GB
Fits in memory315 tok/sest · 262K ctx
Fits in memory269 tok/sest · 262K ctx
Fits in memory180 tok/sest · 262K ctx