New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Can the Instinct MI455X run Lumina-Image-2.0?
Yes — here's how fast and up to what context, with real measured numbers.
Lumina-Image-2.0 on the Instinct MI455X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 5962k | 2374.6-2902.3 t/s | ~4.8-5.9 (slow) |
| Q8 | yes | 6002k | 4749.2-5804.6 t/s | ~9.7-11.8 (slow) |
| Q4 | yes | 6019k | 8480.8-10365.4 t/s | ~17.3-21.2 (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 MI455X run Lumina-Image-2.0?▶
Yes. At Q4 it generates ~8480.8-10365.4 tok/s and fits up to 6019k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~17.3-21.2 tok/s.
How much context fits?▶
At Q4, up to about 6019k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.