Models /Phi 4 Multimodal Instruct /Can I run it?

Compatibility check

Can I run Phi 4 Multimodal Instruct on Apple M4 (10-core GPU)?

Yes — it runs.best quant:

Decode speed

25 tok/sest

Usable context

131K

Memory

32 GB

Fits in memory25 tok/sest · 131K ctx

Fits in memory21 tok/sest · 131K ctx

Fits in memory14 tok/sest · 66K ctx

What is quantisation? →

Phi 4 Multimodal Instruct

5.6B · full specs & benchmarks →

Apple M4 (10-core GPU)

32 GB · everything it runs →