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 Retrieval-Augmented Generator?
Yes — here's how fast and up to what context, with real measured numbers.
Retrieval-Augmented Generator on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2653k | 4664.5-5559.1 t/s | ~23.3-27.8 (slow) |
| Q8 | yes | 2663k | 9329.1-11118.2 t/s | ~46.6-55.6 (slow) |
| Q4 | yes | 2667k | 16659.1-19853.9 t/s | ~83.3-99.3 (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 Retrieval-Augmented Generator?▶
Yes. At Q4 it generates ~16659.1-19853.9 tok/s and fits up to 2667k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~83.3-99.3 tok/s.
How much context fits?▶
At Q4, up to about 2667k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.