local intelligence

can it run

Can the AMD RX 6700 XT run gpt-oss 20B MoE?

OFFLOAD PARTIAL

14.2 GB vs 12 GB usable — partial CPU offload, expect large speed loss

12GB · RDNA2 gfx1031 · model file 14GB · 20B MoE

~37–91 tok/s (active-expert estimate)

computed roofline: 384 GB/s × 0.20–0.50 efficiency window / 2.1GB (Q4_K_M) — bands, never points.

MoE: file is 14GB (that decides fit) but only ~3B of experts activate per token (~2.1GB streamed) — band below is the active-expert estimate, optimistic; community MoE rows run several× the naive bandwidth math

Run gpt-oss 20B MoE on instead

Tesla P40 (24GB, used) — cheapest catalog machine that runs gpt-oss 20B MoE fully ($290)

radeon-9070-xt — the next step up from AMD RX 6700 XT

every model the AMD RX 6700 XT 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 RX 6700 XT · fit = weights + 4k KV vs usable memory