🔌
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 Apple M2 Ultra 64GB run Qwen-VLA?
Yes — here's how fast and up to what context, with real measured numbers.
Qwen-VLA on the Apple M2 Ultra 64GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 349k | 52.8-63.7 t/s | ~0.7-0.8 (slow) |
| Q8 | yes | 389k | 105.6-127.4 t/s | ~1.3-1.6 (slow) |
| Q4 | yes | 406k | 188.6-227.5 t/s | ~2.4-2.8 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the Apple M2 Ultra 64GB run Qwen-VLA?▶
Yes. At Q4 it generates ~188.6-227.5 tok/s and fits up to 406k context; beyond that it spills to SSD swap and slows to ~2.4-2.8 tok/s.
How much context fits?▶
At Q4, up to about 406k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.