// RUN โ€” FITS

Can the AMD Instinct MI455X run Grok-1?

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

Grok-1 on the AMD Instinct MI455X

QuantFits?Max contextSpeedBeyond context
FP16 no โ€” โ€” โ€”
Q8 yes 156k 39.3-48.1 t/s ~0.1-0.1 (slow)
Q4 yes 420k 70.2-85.8 t/s ~0.1-0.2 (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 MI455X run Grok-1?โ–ถ

Yes. At Q4 it generates ~70.2-85.8 tok/s and fits up to 420k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~0.1-0.2 tok/s.

How much context fits?โ–ถ

At Q4, up to about 420k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.