🔌
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 StarCoder 2 3B?
Yes — here's how fast and up to what context, with real measured numbers.
StarCoder 2 3B on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2581k | 973.3-1160 t/s | ~4.9-5.8 (slow) |
| Q8 | yes | 2627k | 1946.7-2320 t/s | ~9.7-11.6 (slow) |
| Q4 | yes | 2647k | 3476.2-4142.9 t/s | ~17.4-20.7 (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 B200 run StarCoder 2 3B?▶
Yes. At Q4 it generates ~3476.2-4142.9 tok/s and fits up to 2647k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~17.4-20.7 tok/s.
How much context fits?▶
At Q4, up to about 2647k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.