ResearchBenchmarksAutonomous Agent Task Solving on AgentBoard BabyaiFollow100Task CompletenessAdMem93.44895.14996.8598.551Jun 5, 2026Evaluation ResultsMethodMethodLinksTask CompletenessAverage ProgressAdMemLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.06100100ReActLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.0697.397.32AWMLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.0693.793.682