Compatibility check
Can I run Llama 3 8B Lunaris on Apple M2 Pro (19-core GPU)?
Yes — it runs.best quant:
Decode speed
29 tok/sest
Usable context
8K
Memory
32 GB
Fits in memory29 tok/sest · 8K ctx
Fits in memory25 tok/sest · 8K ctx
Fits in memory17 tok/sest · 8K ctx