| Benchmark | Model | Rank | Results |
|---|---|---|---|
| mathematical-reasoning-on-mathverse | TGRL-DAPO (Qwen2.5-VL-7B, 2.1K) | – | Accuracy: 55.21 |
| multimodal-reasoning-on-mathvision | TGRL-DAPO (Qwen2.5-VL-7B, 39K) | – | Accuracy: 31.26 |
| multimodal-reasoning-on-mathvista | TGRL-DAPO (Qwen2.5-VL-7B, 39K) | – | Accuracy: 75.0 |