// RUN โ€” FITS

Can the Apple M2 Ultra 192GB run Mixtral 8x22B?

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

Mixtral 8x22B on the Apple M2 Ultra 192GB

QuantFits?Max contextSpeedBeyond context
FP16 no โ€” โ€” โ€”
Q8 yes 92k 3.9-4.7 t/s ~0-0.1 (slow)
Q4 yes 282k 6.9-8.3 t/s ~0.1-0.1 (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 192GB run Mixtral 8x22B?โ–ถ

Yes. At Q4 it generates ~6.9-8.3 tok/s and fits up to 282k context; beyond that it spills to SSD swap and slows to ~0.1-0.1 tok/s.

How much context fits?โ–ถ

At Q4, up to about 282k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.