🔌
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 GPT-Neo-1.3B?
Yes — here's how fast and up to what context, with real measured numbers.
GPT-Neo-1.3B on the Apple M2 Ultra 64GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 816k | 209.2-252.3 t/s | ~2.6-3.2 (slow) |
| Q8 | yes | 836k | 418.5-504.6 t/s | ~5.2-6.3 (slow) |
| Q4 | yes | 845k | 747.3-901.1 t/s | ~9.3-11.3 (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 GPT-Neo-1.3B?▶
Yes. At Q4 it generates ~747.3-901.1 tok/s and fits up to 845k context; beyond that it spills to SSD swap and slows to ~9.3-11.3 tok/s.
How much context fits?▶
At Q4, up to about 845k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.