// RUN โ€” FITS

Can the Intel Gaudi 3 run AntiFormer?

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

AntiFormer on the Intel Gaudi 3

QuantFits?Max contextSpeedBeyond context
FP16 yes 1773k 43493.1-53991.4 t/s ~470.2-583.7 (slow)
Q8 yes 1774k 86986.1-107982.8 t/s ~940.4-1167.4 (slow)
Q4 yes 1774k 155332.4-192826.4 t/s ~1679.3-2084.6 (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 AntiFormer?โ–ถ

Yes. At Q4 it generates ~155332.4-192826.4 tok/s and fits up to 1774k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1679.3-2084.6 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1774k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.