🔌
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 Kosmos-2?
Yes — here's how fast and up to what context, with real measured numbers.
Kosmos-2 on the FuriosaAI RNGD
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 602k | 304.7-370.3 t/s | ~8.1-9.9 (slow) |
| Q8 | yes | 627k | 609.4-740.6 t/s | ~16.3-19.8 (slow) |
| Q4 | yes | 637k | 1088.2-1322.5 t/s | ~29-35.3 (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 Kosmos-2?▶
Yes. At Q4 it generates ~1088.2-1322.5 tok/s and fits up to 637k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~29-35.3 tok/s.
How much context fits?▶
At Q4, up to about 637k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.