🔌
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 MI455X run Qwen3-32B?
Yes — here's how fast and up to what context, with real measured numbers.
Qwen3-32B on the Instinct MI455X
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1260k | 188.2-230.1 t/s | ~0.4-0.5 (slow) |
| Q8 | yes | 1385k | 376.5-460.1 t/s | ~0.8-0.9 (slow) |
| Q4 | yes | 1440k | 672.3-821.6 t/s | ~1.4-1.7 (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 Qwen3-32B?▶
Yes. At Q4 it generates ~672.3-821.6 tok/s and fits up to 1440k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~1.4-1.7 tok/s.
How much context fits?▶
At Q4, up to about 1440k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.