| model                          |       size |     params | backend    | ngl | type_k | type_v |  fa |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | --: | --------------: | -------------------: |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           pp512 |      9883.45 ± 91.73 |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |          pp4096 |      9507.75 ± 55.72 |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           tg128 |        239.11 ± 1.01 |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  pp512 @ d16384 |      8642.58 ± 60.21 |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 | pp4096 @ d16384 |      8473.00 ± 18.82 |
| qwen35moe 35B.A3B Q8_0         |  34.36 GiB |    34.66 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  tg128 @ d16384 |        224.59 ± 0.94 |

build: d2462f8f7 (9590)
