Skip to main content
Partial GPU

Best variant: Q4_K_M

Partial GPU offload — 8 GB VRAM is above the 8 GB minimum but below the 9.1 GB recommendation. Some layers will spill to RAM.

GPU VRAM
8 GB
Min VRAM (best fit)
8 GB
Recommended VRAM
9.1 GB
Estimated tok/s
~40.9

Share this matchup

Send this page so a friend can see if Intel Arc A580 fits Phi-4 Reasoning 14B.

Every Phi-4 Reasoning 14B quantization on Intel Arc A580

Each row runs the compatibility engine against your VRAM, RAM, and the model's requirements.

QuantizationFile SizeMin VRAMRec VRAMContextVerdictEstimated tok/s
Q4_K_MBest fit7 GB8 GB9.1 GB8K / 8KPartial GPU~40.9
Q5_K_M8.8 GB10.1 GB11.4 GB8K / 8KHybrid CPU+GPU~17
Q8_014 GB16.1 GB18.2 GB8K / 8KHybrid CPU+GPU~11
FP1628 GB32.2 GB36.4 GB8K / 8KHybrid CPU+GPU~6

Intel Arc A580 is solid pick for Phi-4 Reasoning 14B

Need second card or fresh build? These links help support site at no extra cost.

All hardware for Phi-4 Reasoning 14BBest GPU for Phi-4 Reasoning 14BModels that fit Intel Arc A580Full model detailsBrowse all models