// RUN โ€” FITS

Can the Apple M3 Max 48GB run ADM?

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

ADM on the Apple M3 Max 48GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 619k 243.3-293.4 t/s ~6.1-7.3 (slow)
Q8 yes 628k 486.6-586.8 t/s ~12.2-14.7 (slow)
Q4 yes 632k 868.9-1047.8 t/s ~21.7-26.2 (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 M3 Max 48GB run ADM?โ–ถ

Yes. At Q4 it generates ~868.9-1047.8 tok/s and fits up to 632k context; beyond that it spills to SSD swap and slows to ~21.7-26.2 tok/s.

How much context fits?โ–ถ

At Q4, up to about 632k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.