🔌
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 H100 SXM5 80GB run LateOn-Code?
Yes — here's how fast and up to what context, with real measured numbers.
LateOn-Code on the NVIDIA H100 SXM5 80GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1096k | 9405.8-11209.6 t/s | ~112.3-133.8 (slow) |
| Q8 | yes | 1098k | 18811.5-22419.2 t/s | ~224.6-267.7 (slow) |
| Q4 | yes | 1099k | 33592-40034.3 t/s | ~401.1-478 (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 H100 SXM5 80GB run LateOn-Code?▶
Yes. At Q4 it generates ~33592-40034.3 tok/s and fits up to 1099k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~401.1-478 tok/s.
How much context fits?▶
At Q4, up to about 1099k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.