// RUN โ€” FITS

Can the Intel Gaudi 3 run Baichuan 1-13B?

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

Baichuan 1-13B on the Intel Gaudi 3

QuantFits?Max contextSpeedBeyond context
FP16 yes 548k 80.9-100.4 t/s ~0.9-1.1 (slow)
Q8 yes 629k 161.8-200.8 t/s ~1.7-2.2 (slow)
Q4 yes 664k 288.9-358.6 t/s ~3.1-3.9 (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 Intel Gaudi 3 run Baichuan 1-13B?โ–ถ

Yes. At Q4 it generates ~288.9-358.6 tok/s and fits up to 664k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~3.1-3.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 664k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.