| Benchmark | Model | Rank | Results |
|---|---|---|---|
| multimodal-reasoning-on-mathvision | RLSD (Qwen3-VL-8B-Instruct) | – | Accuracy: 52.73 |
| multimodal-reasoning-on-mathvista | RLSD (Qwen3-VL-8B-Instruct) | – | Accuracy: 78.1 |
| multimodal-reasoning-on-mmmu | RLSD (Qwen3-VL-8B-Instruct) | – | Accuracy: 67.22 |