Maze Navigation on Maze
100RSROptimal
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Optimal2025.06 | 100 | — | 25.224 | |
| PPO2025.06 | 99.64 | — | 27.23 | |
| MPPOvariant=A2025.06 | 99.52 | — | 27.649 | |
| MPPOvariant=B2025.06 | 99.04 | — | 28.104 | |
| Bot A2025.06 | 92.4 | — | 53.45 | |
| Bot B2025.06 | 89 | — | 55.714 | |
| GPT-4+CoTParameters=≈1T2026.05 | 71.8 | 6.1 | — | |
| S-AI-RecursiveParameters=<10M2026.05 | 68.4 | 11.2 | — | |
| TRMParameters=7M2026.05 | 62.3 | 15.4 | — | |
| S-AI-IoTParameters=2M2026.05 | 51.2 | — | — | |
| GPT-3.5Parameters=175B2026.05 | 38.6 | — | — |