List your company

Embodied Planning on ALFWorld Text N=134 unseen tasks Misaligned

501,049Token CountNL-PDDL
312,466.481,585,398.492,858,330.54,131,262.51Jul 7, 2026

Evaluation results

MethodLinks
NL-PDDL2026.07501,04991
EvoPlanBackbone=Qwen3-32B2026.071,055,34198.5
Gemini-2.0 FlashPlanner Type=Direct LLM/VLM-based, Backbone=Gemini-2.0 Flash2026.071,251,53315
LLaMA-3.1Planner Type=Direct LLM/VLM-based, Backbone=LLaMA-3.12026.071,310,66415
GPT-4oPlanner Type=Direct LLM/VLM-based, Backbone=GPT-4o2026.071,440,47017
ReActPlanner Type=Reflective LLM-based, In-context Examples=with model2026.074,428,91423
ReActPlanner Type=Reflective LLM-based, In-context Examples=with examples2026.075,215,61279