| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| MM-Star | CoVT | Accuracy62 | 18 | 2mo ago | |
| ScienceQA | Unsilencing Latent Reasoning | Accuracy84.2 | 18 | 2mo ago | |
| Mathematical and General Reasoning Suite Combined | WIST | Overall Accuracy46.7 | 16 | 4mo ago | |
| Aggregate Benchmarks | Qwen2.5-VL-7B + DUPL | Average Accuracy66.9 | 13 | 1mo ago |