Visual puzzle solving on Jigsaw R1 (test)
61.9Accuracy (2x1)GPT-4.1-mini
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPT-4.1-mini2025.08 | 61.9 | 54.5 | 54.8 | 20.3 | 47.88 | |
| BAGEL2025.08 | 51.25 | 50.05 | 52.05 | 9.6 | 40.73 | |
| Uni-CoT2025.08 | 51.15 | 61.46 | 59.15 | 18.64 | 47.6 | |
| Random2025.08 | 50 | 50 | 50 | 12.5 | 40.63 | |
| Qwen2.5-VL-7BParameters=7B2025.08 | 49.4 | 48.7 | 50.6 | 15.8 | 41.12 | |
| InternVL2.5-2BParameters=2B2025.08 | 44.9 | 41.9 | 48.6 | 9.7 | 36.28 |