🔌
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 Phi-4-Multimodal?
Yes — here's how fast and up to what context, with real measured numbers.
Phi-4-Multimodal on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1251k | 521.4-621.4 t/s | ~2.6-3.1 (slow) |
| Q8 | yes | 1293k | 1042.9-1242.9 t/s | ~5.2-6.2 (slow) |
| Q4 | yes | 1312k | 1862.2-2219.4 t/s | ~9.3-11.1 (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 Phi-4-Multimodal?▶
Yes. At Q4 it generates ~1862.2-2219.4 tok/s and fits up to 1312k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~9.3-11.1 tok/s.
How much context fits?▶
At Q4, up to about 1312k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.