🔌
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 Pythia-160m?
Yes — here's how fast and up to what context, with real measured numbers.
Pythia-160m on the Apple M2 Ultra 64GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 851k | 1700-2050 t/s | ~21.3-25.6 (slow) |
| Q8 | yes | 854k | 3400-4100 t/s | ~42.5-51.3 (slow) |
| Q4 | yes | 855k | 6071.4-7321.4 t/s | ~75.9-91.5 (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 Pythia-160m?▶
Yes. At Q4 it generates ~6071.4-7321.4 tok/s and fits up to 855k context; beyond that it spills to SSD swap and slows to ~75.9-91.5 tok/s.
How much context fits?▶
At Q4, up to about 855k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.