local intelligence

can it run

Can the NVIDIA RTX A6000 run mixtral 8x7B MoE?

FULL

26.5 GB needed of 48 GB usable — headroom for context

48GB · Ampere · model file 26GB · 8x7B MoE

~91–150 tok/s (active-expert estimate)

computed roofline: 768 GB/s × 0.46–0.76 efficiency window / 3.9GB (Q4_K_M) — bands, never points.

MoE: file is 26GB (that decides fit) but only ~1.05B of experts activate per token (~3.9GB streamed) — band below is the active-expert estimate, optimistic; community MoE rows run several× the naive bandwidth math

Run mixtral 8x7B MoE on instead

Mac mini M5 Pro (64GB) — cheapest catalog machine that runs mixtral 8x7B MoE fully ($1669)

RTX 4090 (24GB, used) — upgrade path machine

every model the NVIDIA RTX A6000 can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the NVIDIA RTX A6000 · fit = weights + 4k KV vs usable memory