Compatibility check
Can I run Canary Qwen 2.5b on Apple M5 Max (40-core GPU)?
Yes — it runs.best quant:
Decode speed
273 tok/sest
Usable context
262K
Memory
128 GB
Fits in memory273 tok/sest · 262K ctx
Fits in memory232 tok/sest · 262K ctx
Fits in memory156 tok/sest · 262K ctx