| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| MMBench | BAGEL | Accuracy85 | 141 | 18d ago | |
| MMMU | Accuracy72.9 | 65 | 18d ago | ||
| MMVP | GPT-5 | Accuracy86.33 | 33 | 1mo ago | |
| VizWiz (test) | LLaVA-FastV (k=3, r=0.75) | VizWiz Score54.7 | 24 | 1mo ago | |
| CVBench-2D | Jigsaw + CARE | Accuracy77.76 | 22 | 1mo ago | |
| MMT-47 Vision Benchmark | LiMEDoRA | Accuracy78.12 | 17 | 13d ago | |
| RealworldQA | Overall Score75.4 | 17 | 1mo ago | ||
| SEED | LLaVA-1.5-13B | Accuracy68.2 | 15 | 26d ago | |
| LLaVA-W | VIG training + VAR | Score63 | 10 | 1mo ago | |
| LLaVA-Wild | LLaVA-FastV (k=3, r=0.75) | LLaVA-Wild Accuracy74.2 | 8 | 1mo ago | |
| MMBench v1.0 (test) | LLaVA-1.5 13B + VIG training | Accuracy68.67 | 6 | 1mo ago | |
| MMVet v1.0 (test) | LLaVA-1.5 13B + VIG training | Score36.87 | 6 | 1mo ago | |
| LLaVAW v1.0 (test) | LLaVA-1.5 13B + VIG training | Score73.45 | 6 | 1mo ago | |
| Common Visual Understanding Benchmarks (GQA, MMB, MME, POPE, SEED, SQA, VQAv2) | Upper Bound | GQA Score61.1 | 5 | 1mo ago |
