| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Geometry3K (val) | Qwen2.5-VL-3B-Instruct + VEPO | Accuracy44.93 | 21 | 1mo ago | |
| Q-Spatial++ | TIGeR | δ≤2 Score70.3 | 9 | 4mo ago | |
| Geometry3k (test) | DS (Oracle) | Test Score48.11 | 8 | 4mo ago | |
| Geometry3K | GPT-4o + Sketchpad | Geometry Score66.7 | 7 | 4mo ago | |
| GeoEval (val) | CalibRL | Accuracy33.44 | 6 | 4mo ago |