New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA B200 run Pythia-160m?
Yes — here's how fast and up to what context, with real measured numbers.
Pythia-160m on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2668k | 18250-21750 t/s | ~91.3-108.8 (slow) |
| Q8 | yes | 2670k | 36500-43500 t/s | ~182.5-217.5 (slow) |
| Q4 | yes | 2671k | 65178.6-77678.6 t/s | ~325.9-388.4 (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 B200 run Pythia-160m?▶
Yes. At Q4 it generates ~65178.6-77678.6 tok/s and fits up to 2671k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~325.9-388.4 tok/s.
How much context fits?▶
At Q4, up to about 2671k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.