// RUN โ€” FITS

Can the FuriosaAI RNGD run Phi-3.5-MoE?

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

Phi-3.5-MoE on the FuriosaAI RNGD

QuantFits?Max contextSpeedBeyond context
FP16 no โ€” โ€” โ€”
Q8 no โ€” โ€” โ€”
Q4 yes 26k 28.6-34.8 t/s ~0.8-0.9 (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 FuriosaAI RNGD run Phi-3.5-MoE?โ–ถ

Yes. At Q4 it generates ~28.6-34.8 tok/s and fits up to 26k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~0.8-0.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 26k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.