ResearchBenchmarksMulti-robot task planning on MACE-THOR Parallel-IndependentFollow71Success RateH-AIM2.3620.183855.82Jan 16, 2026Evaluation ResultsMethodMethodLinksSuccess RateGoal Completion RateExecution ScoreH-AIMLLM Backbone=GPT-4oLLM Backbone=GPT-4o2026.0171880.76H-AIMLLM Backbone=Claude-3....LLM Backbone=Claude-3.5-Sonnet2026.0157700.77H-AIMLLM Backbone=DeepSeek-...LLM Backbone=DeepSeek-V3.12026.0119381LaMMA-PLLM Backbone=GPT-4oLLM Backbone=GPT-4o2026.0114400.81H-AIMLLM Backbone=Qwen-MaxLLM Backbone=Qwen-Max2026.015331