| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| NaturalBench | Qwen2.5-VL-7B + SRPO | General Score78.62 | 21 | 2mo ago | |
| LogicVista | Qwen2.5-VL-7B + SRPO | Score50.56 | 21 | 2mo ago | |
| MMMU-Pro | Qwen2.5-VL-7B + SRPO | Score38.09 | 21 | 2mo ago | |
| General Multimodal Reasoning Suite MMK12, Geo3K, MathVerse, MathVista, We-Math | PTD-PO | MMK12 Accuracy76.21 | 18 | 1mo ago | |
| EMMA full | OpenAI-o1 | Accuracy45.7 | 14 | 4mo ago | |
| MMStar (val) | Qwen2.5-VL-72B-IT | Accuracy70.8 | 13 | 4mo ago | |
| M3CoT | MiMo-VL-7B-OpenMMReasoner | Pass@1 Accuracy78.21 | 11 | 4mo ago | |
| MMMU (val) | Qwen3-VL-8B-DeepVision | Pass@1 Accuracy71.33 | 11 | 4mo ago | |
| General Benchmarks | MUPO | Top-1 Accuracy57.8 | 6 | 3mo ago |
