ResearchBenchmarksAutonomous Agent Task Solving on AgentBoard Tool-query AcademicFollow94.7Task CompletenessAdMem53.41264.13174.8585.569Jun 5, 2026Evaluation ResultsMethodMethodLinksTask CompletenessAverage ProgressAdMemLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.0694.796.71ReActLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.067083.65AWMLLM backbone=Claude Ha...LLM backbone=Claude Haiku 4.52026.065581.85