🔌
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 FRED-T5-XL?
Yes — here's how fast and up to what context, with real measured numbers.
FRED-T5-XL on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1903k | 1006.9-1200 t/s | ~8.4-10 (slow) |
| Q8 | yes | 1930k | 2013.8-2400 t/s | ~16.8-20 (slow) |
| Q4 | yes | 1942k | 3596.1-4285.7 t/s | ~30-35.7 (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 FRED-T5-XL?▶
Yes. At Q4 it generates ~3596.1-4285.7 tok/s and fits up to 1942k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~30-35.7 tok/s.
How much context fits?▶
At Q4, up to about 1942k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.