| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-humaneval | Safin-1-35B-A3B-SFT | #29 | Pass@1: 90.24 |
| code-generation-on-mbpp | Safin-1-35B-A3B-SFT | #37 | Accuracy: 78.6 |
| language-modelling-on-lambada | Kimi Delta Attention with MARCH | #39 | Accuracy: 52.57 |
| mathematical-reasoning-on-aime-2025 | Safin-1-35B-A3B-SFT | #39 | Accuracy: 80.1 |
| question-answering-on-piqa | Gated DeltaNet with MARCH | #153 | Accuracy: 66.9 |
| common-sense-reasoning-on-winogrande | Gated DeltaNet with MARCH | #159 | Accuracy: 52.6 |