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 GPT-2 (124M)?
Yes — here's how fast and up to what context, with real measured numbers.
GPT-2 (124M) on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1953k | 14129-16838.7 t/s | ~117.7-140.3 (slow) |
| Q8 | yes | 1955k | 28258.1-33677.4 t/s | ~235.5-280.6 (slow) |
| Q4 | yes | 1955k | 50460.8-60138.2 t/s | ~420.5-501.2 (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 GPT-2 (124M)?▶
Yes. At Q4 it generates ~50460.8-60138.2 tok/s and fits up to 1955k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~420.5-501.2 tok/s.
How much context fits?▶
At Q4, up to about 1955k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.