New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA H200 run Pythia-70m?
Yes — here's how fast and up to what context, with real measured numbers.
Pythia-70m on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1954k | 25028.6-29828.6 t/s | ~208.6-248.6 (slow) |
| Q8 | yes | 1955k | 50057.1-59657.1 t/s | ~417.1-497.1 (slow) |
| Q4 | yes | 1956k | 89387.8-106530.6 t/s | ~744.9-887.8 (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 Pythia-70m?▶
Yes. At Q4 it generates ~89387.8-106530.6 tok/s and fits up to 1956k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~744.9-887.8 tok/s.
How much context fits?▶
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.