Plan Generation on 6 grid world domains Seen Appearances
95.2Success Rate (Frozenlake)VLMFP
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| VLMFPPlanning Approach=VLMFP, VLM Backbone=GPT-4o2025.10 | 95.2 | 88.7 | 55.8 | 75.2 | 58.7 | 46.2 | 70 | |
| CodePDDLPlanning Approach=CodePDDL, VLM Backbone=GPT-4o2025.10 | 88.1 | 87.5 | 0 | 0 | 7.2 | 1.1 | 30.7 | |
| CoTPlanning Approach=Chain-of-Thought, VLM Backbone=GPT-52025.10 | 36 | 14 | 14 | 2 | 10 | 0 | 12.7 | |
| DirectPlanning Approach=Direct, VLM Backbone=GPT-52025.10 | 30 | 10 | 15 | 5 | 11 | 0 | 11.8 | |
| CoTPlanning Approach=Chain-of-Thought, VLM Backbone=GPT-4o2025.10 | 9 | 1 | 0 | 0 | 0 | 0 | 1.7 | |
| DirectPlanning Approach=Direct, VLM Backbone=GPT-4o2025.10 | 7 | 1 | 0 | 0 | 0 | 0 | 1.3 |