Feasibility Prediction on Sokoban
46.4F1@1Claude Sonnet 4.6
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Claude Sonnet 4.62026.05 | 46.4 | 53.6 | 33.9 | |
| Claude Opus 4.72026.05 | 46.3 | 45.6 | 16 | |
| Gemini 3.1 Pro Preview2026.05 | 40 | 61.9 | 79.9 | |
| GPT-5.2 Instant2026.05 | 27.7 | 40.6 | 32.8 | |
| Qwen3 235B2026.05 | 6.3 | 12.6 | 20.4 |