Q: How fast is gpt-oss 120B MoE on the NVIDIA H100 80GB SXM?

~160–260 tok/s (active-expert estimate) on the NVIDIA H100 80GB SXM. computed roofline: 3350 GB/s × 0.46–0.76 efficiency window / 9.8GB (Q4_K_M) — bands, never points. MoE: file is 65GB (that decides fit) but only ~18B of experts activate per token (~9.8GB streamed) — band below is the active-expert estimate, optimistic; community MoE rows run several× the naive bandwidth math

Question 1

Can the NVIDIA H100 80GB SXM run gpt-oss 120B MoE?

Accepted Answer

gpt-oss 120B MoE needs about 65.28GB working set (weights 65GB + KV cache 0.28GB at 4k context). The NVIDIA H100 80GB SXM offers ~80GB usable — verdict: tight. 65.3 of 80 GB usable — keep context modest

Question 2

How fast is gpt-oss 120B MoE on the NVIDIA H100 80GB SXM?

Accepted Answer

~160–260 tok/s (active-expert estimate)

computed roofline: 3350 GB/s × 0.46–0.76 efficiency window / 9.8GB (Q4_K_M) — bands, never points.

MoE: file is 65GB (that decides fit) but only ~18B of experts activate per token (~9.8GB 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 120B MoE on instead?

Accepted Answer

NIMO AI Mini PC (128GB) runs it fully ($1999).

Can the NVIDIA H100 80GB SXM run gpt-oss 120B MoE?

Run gpt-oss 120B MoE on instead