ResearchBenchmarksMulti-robot collaboration on Install Drywall 1.0 (test)Follow100Task Success RateREFLEX60.4870.748191.26May 20, 2025Evaluation ResultsMethodMethodLinksTask Success RateEnvironment StepsReplan AttemptsREFLEXBase LLM=GPT-4Base LLM=GPT-42025.051004.10REFLEXBase LLM=LLaMA-3.1Base LLM=LLaMA-3.12025.05954.21.9RoCoBase LLM=GPT-4Base LLM=GPT-42025.05625.35.8