Spatial Reasoning on Maze 10x10
61.37CR (%)HSRL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HSRL2026.05 | 61.37 | 47.87 | |
| System-1.x2026.05 | 54.74 | 36.02 | |
| Inner Monologue2026.05 | 54.03 | 34.6 | |
| ReAct2026.05 | 53.8 | 26.06 | |
| Reflexion2026.05 | 45.02 | 37.91 | |
| CoT2026.05 | 43.12 | 38.39 | |
| Tree Planner2026.05 | 39.1 | 27.01 | |
| HyperTree2026.05 | 37.91 | 22.98 | |
| ProgPrompt2026.05 | 34.6 | 33.41 | |
| Direct Answer2026.05 | 23.69 | 23.45 |