| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-minerva-math | Qwen3-4B-Base (with OLR) | #17 | Accuracy: 44.9 |
| mathematical-reasoning-on-olympiadbench | Qwen3-4B-Base (with OLR) | #28 | Accuracy: 48.9 |
| mathematical-reasoning-on-math-500 | Qwen3-4B-Base (with OLR) | #70 | Accuracy: 85.2 |
| multi-task-language-understanding-on-mmlu-pro | Qwen3-4B-Base (with OLR) | #95 | Accuracy: 60.6 |
| mathematical-reasoning-on-aime24 | Qwen3-4B-Base (with OLR) | #111 | Acc: 23.8 |
| mathematical-reasoning-on-aime-2025 | Qwen3-4B-Base (with OLR) | #126 | Accuracy: 20.8 |