// RUN โ€” FITS

Can the NVIDIA H200 run Flan UL2?

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

Flan UL2 on the NVIDIA H200

QuantFits?Max contextSpeedBeyond context
FP16 yes 340k 89.8-107.1 t/s ~0.7-0.9 (slow)
Q8 yes 415k 179.7-214.2 t/s ~1.5-1.8 (slow)
Q4 yes 447k 320.9-382.4 t/s ~2.7-3.2 (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 NVIDIA H200 run Flan UL2?โ–ถ

Yes. At Q4 it generates ~320.9-382.4 tok/s and fits up to 447k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~2.7-3.2 tok/s.

How much context fits?โ–ถ

At Q4, up to about 447k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.