Geometric problem solving on GeoAux
42.97AccuracyA²PO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| A²POBase Model=Qwen2.5-VL-7B-Instruct2026.03 | 42.97 | 1.0869 | 55.76 | |
| GeometryZeroBase Model=Qwen2.5-VL-7B-Instruct2026.03 | 40.18 | 1.0945 | 54.07 | |
| ToRLBase Model=Qwen2.5-VL-7B-Instruct2026.03 | 39.77 | 1.0941 | 52.92 | |
| GRPOBase Model=Qwen2.5-VL-7B-Instruct2026.03 | 39.28 | 1.0887 | 54.06 | |
| SFTBase Model=Qwen2.5-VL-7B-Instruct2026.03 | 37.3 | 1.0857 | 49.02 | |
| A²POBase Model=Qwen2.5-VL-3B-Instruct2026.03 | 33.2 | 1.1534 | 48.22 | |
| GRPOBase Model=Qwen2.5-VL-3B-Instruct2026.03 | 31.22 | 1.155 | 47.01 | |
| GeometryZeroBase Model=Qwen2.5-VL-3B-Instruct2026.03 | 29.33 | 1.1535 | 46.35 | |
| ToRLBase Model=Qwen2.5-VL-3B-Instruct2026.03 | 28.68 | 1.1558 | 44.71 | |
| SFTBase Model=Qwen2.5-VL-3B-Instruct2026.03 | 23.09 | 1.1389 | 39.56 |