Models /Granite 4.0 Micro /Can I run it?

Compatibility check

Can I run Granite 4.0 Micro on Apple M5 Max (40-core GPU)?

Yes — it runs.best quant:

Decode speed

222 tok/sest

Usable context

66K

Memory

128 GB

Fits in memory222 tok/sest · 66K ctx

Fits in memory189 tok/sest · 66K ctx

Fits in memory126 tok/sest · 66K ctx

What is quantisation? →

Granite 4.0 Micro

3.2B · full specs & benchmarks →

Apple M5 Max (40-core GPU)

128 GB · everything it runs →