🔌
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 H200 run VILA-13B?
Yes — here's how fast and up to what context, with real measured numbers.
VILA-13B on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 620k | 131.2-156.4 t/s | ~1.1-1.3 (slow) |
| Q8 | yes | 701k | 262.5-312.8 t/s | ~2.2-2.6 (slow) |
| Q4 | yes | 737k | 468.7-558.6 t/s | ~3.9-4.7 (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 H200 run VILA-13B?▶
Yes. At Q4 it generates ~468.7-558.6 tok/s and fits up to 737k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~3.9-4.7 tok/s.
How much context fits?▶
At Q4, up to about 737k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.