// RUN โ€” FITS

Can the NVIDIA H200 run RWKV-4 World (7B)?

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

RWKV-4 World (7B) on the NVIDIA H200

QuantFits?Max contextSpeedBeyond context
FP16 yes 865k 237-282.4 t/s ~2-2.4 (slow)
Q8 yes 922k 474-564.9 t/s ~3.9-4.7 (slow)
Q4 yes 947k 846.4-1008.7 t/s ~7.1-8.4 (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 RWKV-4 World (7B)?โ–ถ

Yes. At Q4 it generates ~846.4-1008.7 tok/s and fits up to 947k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~7.1-8.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 947k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.