// RUN โ€” FITS

Can the Apple M2 Ultra 64GB run ChemBERTa?

Yes โ€” here's how fast and up to what context, with real measured numbers.

ChemBERTa on the Apple M2 Ultra 64GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 852k 2176-2624 t/s ~27.2-32.8 (slow)
Q8 yes 854k 4352-5248 t/s ~54.4-65.6 (slow)
Q4 yes 855k 7771.4-9371.4 t/s ~97.1-117.1 (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 ChemBERTa?โ–ถ

Yes. At Q4 it generates ~7771.4-9371.4 tok/s and fits up to 855k context; beyond that it spills to SSD swap and slows to ~97.1-117.1 tok/s.

How much context fits?โ–ถ

At Q4, up to about 855k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.