| mathematical-reasoning-on-aime-2025 | Qwen2.5-Math-7B + GRPO (P2T token reward) (extra data) | – | Accuracy: 20.0 |
| mathematical-reasoning-on-aime24 | Qwen2.5-Math-7B + GRPO (P2T token reward) (extra data) | – | Acc: 40.3 |
| mathematical-reasoning-on-math-500 | Qwen2.5-Math-7B + GRPO (P2T token reward) (extra data) | – | Accuracy: 84.5 |
| mathematical-reasoning-on-mathverse | Qwen2.5-VL-7B-Instruct + GRPO (P2T token reward) (extra data) | – | Accuracy: 51.0 |
| mathematical-reasoning-on-minerva-math | Qwen2.5-Math-7B + GRPO (P2T token reward) (extra data) | – | Accuracy: 41.9 |
| multimodal-reasoning-on-mathvision | Qwen2.5-VL-7B-Instruct + GRPO (P2T token reward) (extra data) | – | Accuracy: 30.9 |
| multimodal-reasoning-on-mathvista | Qwen2.5-VL-7B-Instruct + GRPO (P2T token reward) (extra data) | – | Accuracy: 75.0 |
| multimodal-reasoning-on-mmmu | Qwen2.5-VL-7B-Instruct + GRPO (P2T token reward) (extra data) | – | Accuracy: 63.4 |
| vision-language-understanding-on-mmstar | Qwen2.5-VL-7B-Instruct + GRPO (P2T token reward) (extra data) | – | Accuracy: 67.2 |