// RUN โ€” FITS

Can the Apple M1 Ultra 128GB run LongT5?

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

LongT5 on the Apple M1 Ultra 128GB

QuantFits?Max contextSpeedBeyond context
FP16 yes 1643k 90.7-109.3 t/s ~1.1-1.4 (slow)
Q8 yes 1689k 181.3-218.7 t/s ~2.3-2.7 (slow)
Q4 yes 1709k 323.8-390.5 t/s ~4-4.9 (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 M1 Ultra 128GB run LongT5?โ–ถ

Yes. At Q4 it generates ~323.8-390.5 tok/s and fits up to 1709k context; beyond that it spills to SSD swap and slows to ~4-4.9 tok/s.

How much context fits?โ–ถ

At Q4, up to about 1709k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.