| model                          |       size |     params | backend    | ngl | type_k | type_v |  fa |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | --: | --------------: | -------------------: |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           pp512 |      9498.56 ± 53.54 |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |          pp4096 |      9446.30 ± 49.06 |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           tg128 |        275.02 ± 0.85 |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  pp512 @ d16384 |      7847.78 ± 47.57 |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 | pp4096 @ d16384 |      7843.39 ± 13.93 |
| gpt-oss 120B MXFP4 MoE         |  59.02 GiB |   116.83 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  tg128 @ d16384 |        235.05 ± 0.56 |

build: d2462f8f7 (9590)