🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the NVIDIA B200 run P-LLama3?
Yes — here's how fast and up to what context, with real measured numbers.
P-LLama3 on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1214k | 365-435 t/s | ~1.8-2.2 (slow) |
| Q8 | yes | 1275k | 730-870 t/s | ~3.7-4.4 (slow) |
| Q4 | yes | 1302k | 1303.6-1553.6 t/s | ~6.5-7.8 (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 B200 run P-LLama3?▶
Yes. At Q4 it generates ~1303.6-1553.6 tok/s and fits up to 1302k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~6.5-7.8 tok/s.
How much context fits?▶
At Q4, up to about 1302k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.