// RUN โ€” FITS

Can the Intel Arc Pro B70 run SANA 1.5 4.8B?

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

SANA 1.5 4.8B on the Intel Arc Pro B70

QuantFits?Max contextSpeedBeyond context
FP16 yes 280k 36.7-45.6 t/s ~2.4-3 (slow)
Q8 yes 353k 73.5-91.2 t/s ~4.8-6 (slow)
Q4 yes 385k 131.2-162.9 t/s ~8.6-10.7 (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 B70 run SANA 1.5 4.8B?โ–ถ

Yes. At Q4 it generates ~131.2-162.9 tok/s and fits up to 385k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~8.6-10.7 tok/s.

How much context fits?โ–ถ

At Q4, up to about 385k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.