| model                          |       size |     params | backend    | ngl | type_k | type_v |  fa |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | --: | --------------: | -------------------: |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           pp512 |     1906.94 ± 220.54 |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |          pp4096 |      2529.59 ± 14.31 |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           tg128 |         83.84 ± 0.24 |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  pp512 @ d16384 |       1432.56 ± 7.23 |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 | pp4096 @ d16384 |       1712.34 ± 4.76 |
| qwen3moe 235B.A22B Q4_K - Medium | 132.39 GiB |   235.09 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  tg128 @ d16384 |         53.55 ± 0.15 |

build: d2462f8f7 (9590)
