local intelligence

can it run

Can the Apple M4 run devstral 24B?

OFFLOAD PARTIAL

15.2 GB vs 14 GB usable — partial CPU offload, expect large speed loss With a q8 (quantized) KV cache it comes to the edge (14.9GB) — possible with modest context, at some quality cost.

~14GB usable · Apple M4 · model file 14.6GB · 24B

~6.6–8.2 tok/s

computed roofline: 120 GB/s × 0.80–1.00 efficiency window / 14.6GB (Q4_K_M) — bands, never points.

Run devstral 24B on instead

Tesla P40 (24GB, used) — cheapest catalog machine that runs devstral 24B fully ($290)

every model the Apple M4 can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the Apple M4 · fit = weights + 4k KV vs usable memory