// RUN โ€” FITS

Can the Intel Arc Pro B50 run Phi-4-Reasoning?

Yes โ€” here's how fast and up to what context, with real measured numbers.

Phi-4-Reasoning on the Intel Arc Pro B50

QuantFits?Max contextSpeedBeyond context
FP16 no โ€” โ€” โ€”
Q8 no โ€” โ€” โ€”
Q4 yes 33k 16.6-20.6 t/s ~3-3.7 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.

FAQ

Can the Intel Arc Pro B50 run Phi-4-Reasoning?โ–ถ

Yes. At Q4 it generates ~16.6-20.6 tok/s and fits up to 33k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~3-3.7 tok/s.

How much context fits?โ–ถ

At Q4, up to about 33k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.