local intelligence

can it run

Can the AMD Ryzen AI Max (Strix Halo) run FLUX dev fp8?

TIGHT

25.0 of 30 GB usable — keep context modest

~30GB usable · RDNA3.5 (Strix Halo) · model file 23GB · dev fp8

~6.7–9.5 tok/s

computed roofline: 256 GB/s × 0.60–0.85 efficiency window / 23.0GB (Q4_K_M) — bands, never points.

Run FLUX dev fp8 on instead

Mac mini M5 Pro (64GB) — cheapest catalog machine that runs FLUX dev fp8 fully ($1669)

every model the AMD Ryzen AI Max (Strix Halo) can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the AMD Ryzen AI Max (Strix Halo) · fit = weights + 4k KV vs usable memory