Compatibility check
Can I run Qwen3 ASR 1.7B HF on Apple M5 Max (40-core GPU)?
Yes — it runs.best quant:
Decode speed
355 tok/sest
Usable context
262K
Memory
128 GB
Fits in memory355 tok/sest · 262K ctx
Fits in memory302 tok/sest · 262K ctx
Fits in memory202 tok/sest · 262K ctx