🔌
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 Gemma 2 2B?
Yes — here's how fast and up to what context, with real measured numbers.
Gemma 2 2B on the NVIDIA B200
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2593k | 1123.1-1338.5 t/s | ~5.6-6.7 (slow) |
| Q8 | yes | 2633k | 2246.2-2676.9 t/s | ~11.2-13.4 (slow) |
| Q4 | yes | 2650k | 4011-4780.2 t/s | ~20.1-23.9 (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 Gemma 2 2B?▶
Yes. At Q4 it generates ~4011-4780.2 tok/s and fits up to 2650k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~20.1-23.9 tok/s.
How much context fits?▶
At Q4, up to about 2650k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.