New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the NVIDIA H200 run LFM2.5-ColBERT-350M?
Yes — here's how fast and up to what context, with real measured numbers.
LFM2.5-ColBERT-350M on the NVIDIA H200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1946k | 5005.7-5965.7 t/s | ~41.7-49.7 (slow) |
| Q8 | yes | 1951k | 10011.4-11931.4 t/s | ~83.4-99.4 (slow) |
| Q4 | yes | 1953k | 17877.6-21306.1 t/s | ~149-177.6 (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 LFM2.5-ColBERT-350M?▶
Yes. At Q4 it generates ~17877.6-21306.1 tok/s and fits up to 1953k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~149-177.6 tok/s.
How much context fits?▶
At Q4, up to about 1953k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.