🔌
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 Intel Gaudi 3 run Qwen-VLA?
Yes — here's how fast and up to what context, with real measured numbers.
Qwen-VLA on the Intel Gaudi 3
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 808k | 208.3-258.6 t/s | ~2.3-2.8 (slow) |
| Q8 | yes | 848k | 416.7-517.3 t/s | ~4.5-5.6 (slow) |
| Q4 | yes | 865k | 744.1-923.7 t/s | ~8-10 (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 Intel Gaudi 3 run Qwen-VLA?▶
Yes. At Q4 it generates ~744.1-923.7 tok/s and fits up to 865k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~8-10 tok/s.
How much context fits?▶
At Q4, up to about 865k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.