🔌
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 96GB run Spec-Drafter?
Yes — here's how fast and up to what context, with real measured numbers.
Spec-Drafter on the Apple M2 Max 96GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1280k | 272-328 t/s | ~6.8-8.2 (slow) |
| Q8 | yes | 1288k | 544-656 t/s | ~13.6-16.4 (slow) |
| Q4 | yes | 1291k | 971.4-1171.4 t/s | ~24.3-29.3 (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 Max 96GB run Spec-Drafter?▶
Yes. At Q4 it generates ~971.4-1171.4 tok/s and fits up to 1291k context; beyond that it spills to SSD swap and slows to ~24.3-29.3 tok/s.
How much context fits?▶
At Q4, up to about 1291k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.