| model                          |       size |     params | backend    | ngl | type_k | type_v |  fa |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | --: | --------------: | -------------------: |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           pp512 |     1785.69 ± 579.74 |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |          pp4096 |       3263.38 ± 9.08 |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |           tg128 |        129.57 ± 0.18 |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  pp512 @ d16384 |       1783.55 ± 5.66 |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 | pp4096 @ d16384 |       2404.90 ± 5.72 |
| minimax-m2 230B.A10B Q4_K - Medium | 131.10 GiB |   228.69 B | CUDA       | 999 |   q8_0 |   q8_0 |   1 |  tg128 @ d16384 |         89.74 ± 0.10 |

build: d2462f8f7 (9590)
