| Benchmark | Model | Rank | Results |
|---|---|---|---|
| instruction-following-on-ifbench | GEPO (Qwen3.5-9B) | – | Accuracy: 80.7 |
| mathematical-reasoning-on-aime-2025 | GEPO (Qwen3.5-9B) | – | Accuracy: 91.4 |
| multimodal-reasoning-on-mathvista | GEPO (Qwen3.5-9B) | – | Accuracy: 84.9 |
| multimodal-reasoning-on-mmmu-pro | GEPO (Qwen3.5-9B) | – | Accuracy: 76.6 |