🔌
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 192GB run mBART-50?
Yes — here's how fast and up to what context, with real measured numbers.
mBART-50 on the Apple M2 Ultra 192GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 2595k | 445.9-537.7 t/s | ~5.6-6.7 (slow) |
| Q8 | yes | 2605k | 891.8-1075.4 t/s | ~11.1-13.4 (slow) |
| Q4 | yes | 2609k | 1592.5-1920.4 t/s | ~19.9-24 (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 192GB run mBART-50?▶
Yes. At Q4 it generates ~1592.5-1920.4 tok/s and fits up to 2609k context; beyond that it spills to SSD swap and slows to ~19.9-24 tok/s.
How much context fits?▶
At Q4, up to about 2609k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.