// RUN โ€” FITS

Can the Apple M2 Max 96GB run ProtBERT-BFD?

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

ProtBERT-BFD on the Apple M2 Max 96GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 1283k 323.8-390.5 t/s ~8.1-9.8 (slow)
Q8 yes 1289k 647.6-781 t/s ~16.2-19.5 (slow)
Q4 yes 1292k 1156.5-1394.6 t/s ~28.9-34.9 (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 ProtBERT-BFD?โ–ถ

Yes. At Q4 it generates ~1156.5-1394.6 tok/s and fits up to 1292k context; beyond that it spills to SSD swap and slows to ~28.9-34.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1292k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.