🔌
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 SeaLLM-7B-v2.5?
Yes — here's how fast and up to what context, with real measured numbers.
SeaLLM-7B-v2.5 on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 848k | 205.2-244.5 t/s | ~1.7-2 (slow) |
| Q8 | yes | 913k | 410.3-489 t/s | ~3.4-4.1 (slow) |
| Q4 | yes | 942k | 732.7-873.2 t/s | ~6.1-7.3 (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 SeaLLM-7B-v2.5?▶
Yes. At Q4 it generates ~732.7-873.2 tok/s and fits up to 942k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~6.1-7.3 tok/s.
How much context fits?▶
At Q4, up to about 942k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.