// RUN โ€” FITS

Can the NVIDIA H200 run ESM1-43M?

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

ESM1-43M on the NVIDIA H200

QuantFits?Max contextSpeedBeyond context
FP16 yes 1955k 41126.8-49014.1 t/s ~342.7-408.5 (slow)
Q8 yes 1956k 82253.5-98028.2 t/s ~685.4-816.9 (slow)
Q4 yes 1956k 146881.3-175050.3 t/s ~1224-1458.8 (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 ESM1-43M?โ–ถ

Yes. At Q4 it generates ~146881.3-175050.3 tok/s and fits up to 1956k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1224-1458.8 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1956k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.