🔌
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 InternVL2_5-26B?
Yes — here's how fast and up to what context, with real measured numbers.
InternVL2_5-26B on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 295k | 68.7-81.9 t/s | ~0.6-0.7 (slow) |
| Q8 | yes | 392k | 137.4-163.8 t/s | ~1.1-1.4 (slow) |
| Q4 | yes | 435k | 245.4-292.4 t/s | ~2-2.4 (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 InternVL2_5-26B?▶
Yes. At Q4 it generates ~245.4-292.4 tok/s and fits up to 435k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~2-2.4 tok/s.
How much context fits?▶
At Q4, up to about 435k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.