🔌
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 DeepHermes 3 - Mistral 24B?
Yes — here's how fast and up to what context, with real measured numbers.
DeepHermes 3 - Mistral 24B on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 485k | 121.7-145 t/s | ~0.6-0.7 (slow) |
| Q8 | yes | 577k | 243.3-290 t/s | ~1.2-1.5 (slow) |
| Q4 | yes | 617k | 434.5-517.9 t/s | ~2.2-2.6 (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 DeepHermes 3 - Mistral 24B?▶
Yes. At Q4 it generates ~434.5-517.9 tok/s and fits up to 617k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~2.2-2.6 tok/s.
How much context fits?▶
At Q4, up to about 617k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.