Geometry Problem Solving on GeoQA (test)
92.3Choice AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| Human2025.05 | 92.3 | |
| AutoGPSBackbone=InternVL32025.05 | 86.2 | |
| FGeo-HyperGNet2025.05 | 85.6 | |
| DualGeoSolver2025.05 | 65.2 | |
| SCA-GPS2025.05 | 64.1 | |
| CalibRL2026.02 | 60.74 | |
| NGS-Auxiliary2025.05 | 60 | |
| NGS2025.05 | 57.4 | |
| GRPO2026.02 | 52.52 | |
| DAPO2026.02 | 52.52 | |
| LUFFY2026.02 | 51.19 | |
| RL-PLUS2026.02 | 46.42 | |
| SFT+GRPO2026.02 | 40.98 |