New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the FuriosaAI RNGD run GPT-2 (124M)?
Yes — here's how fast and up to what context, with real measured numbers.
GPT-2 (124M) on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 647k | 3931.5-4778.2 t/s | ~104.8-127.4 (slow) |
| Q8 | yes | 649k | 7862.9-9556.5 t/s | ~209.7-254.8 (slow) |
| Q4 | yes | 650k | 14040.9-17065.1 t/s | ~374.4-455.1 (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 FuriosaAI RNGD run GPT-2 (124M)?▶
Yes. At Q4 it generates ~14040.9-17065.1 tok/s and fits up to 650k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~374.4-455.1 tok/s.
How much context fits?▶
At Q4, up to about 650k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.