Compatibility check
Can I run Qwen3 ASR 1.7B on Apple M5 Max (40-core GPU)?
Yes — it runs.best quant:
Decode speed
308 tok/sest
Usable context
262K
Memory
128 GB
Fits in memory308 tok/sest · 262K ctx
Fits in memory263 tok/sest · 262K ctx
Fits in memory176 tok/sest · 262K ctx