🔌
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 ALBERT-xxlarge?
Yes — here's how fast and up to what context, with real measured numbers.
ALBERT-xxlarge on the Apple M2 Max 96GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1288k | 578.7-697.9 t/s | ~14.5-17.4 (slow) |
| Q8 | yes | 1292k | 1157.4-1395.7 t/s | ~28.9-34.9 (slow) |
| Q4 | yes | 1293k | 2066.9-2492.4 t/s | ~51.7-62.3 (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 ALBERT-xxlarge?▶
Yes. At Q4 it generates ~2066.9-2492.4 tok/s and fits up to 1293k context; beyond that it spills to SSD swap and slows to ~51.7-62.3 tok/s.
How much context fits?▶
At Q4, up to about 1293k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.