🔌
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 Max 64GB run GPT-2 Medium (FlashAttention)?
Yes — here's how fast and up to what context, with real measured numbers.
GPT-2 Medium (FlashAttention) on the Apple M2 Max 64GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 845k | 383.1-462 t/s | ~9.6-11.5 (slow) |
| Q8 | yes | 851k | 766.2-923.9 t/s | ~19.2-23.1 (slow) |
| Q4 | yes | 853k | 1368.2-1649.9 t/s | ~34.2-41.2 (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 Max 64GB run GPT-2 Medium (FlashAttention)?▶
Yes. At Q4 it generates ~1368.2-1649.9 tok/s and fits up to 853k context; beyond that it spills to SSD swap and slows to ~34.2-41.2 tok/s.
How much context fits?▶
At Q4, up to about 853k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.