🔌
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 B200 run Tranception?
Yes — here's how fast and up to what context, with real measured numbers.
Tranception on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2651k | 4171.4-4971.4 t/s | ~20.9-24.9 (slow) |
| Q8 | yes | 2662k | 8342.9-9942.9 t/s | ~41.7-49.7 (slow) |
| Q4 | yes | 2666k | 14898-17755.1 t/s | ~74.5-88.8 (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 B200 run Tranception?▶
Yes. At Q4 it generates ~14898-17755.1 tok/s and fits up to 2666k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~74.5-88.8 tok/s.
How much context fits?▶
At Q4, up to about 2666k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.