🔌
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 TinyLlama-1.1B (3T token checkpoint)?
Yes — here's how fast and up to what context, with real measured numbers.
TinyLlama-1.1B (3T token checkpoint) on the Apple M2 Max 64GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 822k | 123.6-149.1 t/s | ~3.1-3.7 (slow) |
| Q8 | yes | 839k | 247.3-298.2 t/s | ~6.2-7.5 (slow) |
| Q4 | yes | 847k | 441.6-532.5 t/s | ~11-13.3 (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 64GB run TinyLlama-1.1B (3T token checkpoint)?▶
Yes. At Q4 it generates ~441.6-532.5 tok/s and fits up to 847k context; beyond that it spills to SSD swap and slows to ~11-13.3 tok/s.
How much context fits?▶
At Q4, up to about 847k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.