🔌
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 Instinct MI440X run Qwen2.5-Omni 3B?
Yes — here's how fast and up to what context, with real measured numbers.
Qwen2.5-Omni 3B on the Instinct MI440X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 5950k | 2058-2515.3 t/s | ~4.2-5.1 (slow) |
| Q8 | yes | 5996k | 4116-5030.7 t/s | ~8.4-10.3 (slow) |
| Q4 | yes | 6016k | 7350-8983.3 t/s | ~15-18.3 (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 Instinct MI440X run Qwen2.5-Omni 3B?▶
Yes. At Q4 it generates ~7350-8983.3 tok/s and fits up to 6016k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~15-18.3 tok/s.
How much context fits?▶
At Q4, up to about 6016k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.