| model                          |       size |     params | backend    | ngl | type_k | type_v |  fa |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | --: | --------------: | -------------------: |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           pp512 |      907.08 ± 132.10 |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |          pp4096 |       1407.98 ± 3.51 |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           tg128 |         73.38 ± 0.08 |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  pp512 @ d16384 |        837.25 ± 6.42 |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 | pp4096 @ d16384 |       1090.53 ± 5.28 |
| qwen3moe ?B Q2_K - Medium      | 167.91 GiB |   480.15 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  tg128 @ d16384 |         48.92 ± 0.08 |

build: d2462f8f7 (9590)
