// RUN โ€” FITS

Can the AMD Instinct MI325X run Mistral Medium 3.5?

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

Mistral Medium 3.5 on the AMD Instinct MI325X

QuantFits?Max contextSpeedBeyond context
FP16 no โ€” โ€” โ€”
Q8 yes 324k 29.5-36.1 t/s ~0.2-0.2 (slow)
Q4 yes 495k 52.7-64.5 t/s ~0.4-0.4 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.

FAQ

Can the AMD Instinct MI325X run Mistral Medium 3.5?โ–ถ

Yes. At Q4 it generates ~52.7-64.5 tok/s and fits up to 495k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~0.4-0.4 tok/s.

How much context fits?โ–ถ

At Q4, up to about 495k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.