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

Compatibility check

Can I run Phi 4 Multimodal Instruct on GeForce RTX 5070?

Yes — it runs.best quant:

Decode speed

158 tok/sest

Usable context

33K

Memory

12 GB

Fits in memory158 tok/sest · 33K ctx

Fits in memory135 tok/sest · 33K ctx

Fits in memory90 tok/sest · 16K ctx

What is quantisation? →

Phi 4 Multimodal Instruct

5.6B · full specs & benchmarks →

GeForce RTX 5070

12 GB · everything it runs →