Compatibility check
Can I run Qwen3 ASR 0.6B HF on Apple M3 Ultra (80-core GPU)?
Yes — it runs.best quant:
Decode speed
1182 tok/sest
Usable context
262K
Memory
512 GB
Fits in memory1182 tok/sest · 262K ctx
Fits in memory1008 tok/sest · 262K ctx
Fits in memory675 tok/sest · 262K ctx