Q: How fast is gpt-oss 20B MoE on the AMD Radeon RX 9070 (non-XT)?

~170–250 tok/s (active-expert estimate) on the AMD Radeon RX 9070 (non-XT). computed roofline: 640 GB/s × 0.55–0.82 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

Question 1

Can the AMD Radeon RX 9070 (non-XT) run gpt-oss 20B MoE?

Accepted Answer

gpt-oss 20B MoE needs about 14.19GB working set (weights 14GB + KV cache 0.19GB at 4k context). The AMD Radeon RX 9070 (non-XT) offers ~16GB usable — verdict: tight. 14.2 of 16 GB usable — keep context modest

Question 2

How fast is gpt-oss 20B MoE on the AMD Radeon RX 9070 (non-XT)?

Accepted Answer

~170–250 tok/s (active-expert estimate)

computed roofline: 640 GB/s × 0.55–0.82 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

Question 3

What should I run gpt-oss 20B MoE on instead?

Accepted Answer

Tesla P40 (24GB, used) runs it fully ($290).

Can the AMD Radeon RX 9070 (non-XT) run gpt-oss 20B MoE?

Run gpt-oss 20B MoE on instead