| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| SugarCrepe | DINOv2 | Accuracy88.06 | 20 | 4mo ago | |
| VisMin Hard (test) | Qwen2.5-VL-7B + FOCUS | Text Score (T)84.47 | 16 | 1mo ago | |
| ColorSwap | TTM | Accuracy85.17 | 3 | 3mo ago | |
| WhatsUp B (1 × 4 groups) | TTM | GroupScore49.94 | 2 | 3mo ago | |
| WhatsUp A 1 × 4 groups | TTM | GroupScore56.8 | 2 | 3mo ago |