Compatibility check
Can I run Gemma 4 26B A4B on Apple M4 Max (40-core GPU)?
Yes — it runs.best quant:
Decode speed
134 tok/sest
Usable context
262K
Memory
128 GB
Fits in memory134 tok/sest · 262K ctx
Fits in memory114 tok/sest · 262K ctx
Fits in memory76 tok/sest · 262K ctx