🔌
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 FuriosaAI RNGD run GBERT-Large?
Yes — here's how fast and up to what context, with real measured numbers.
GBERT-Large on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 641k | 1455.2-1768.7 t/s | ~38.8-47.2 (slow) |
| Q8 | yes | 646k | 2910.4-3537.3 t/s | ~77.6-94.3 (slow) |
| Q4 | yes | 648k | 5197.2-6316.6 t/s | ~138.6-168.4 (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 GBERT-Large?▶
Yes. At Q4 it generates ~5197.2-6316.6 tok/s and fits up to 648k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~138.6-168.4 tok/s.
How much context fits?▶
At Q4, up to about 648k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.