ResearchBenchmarksMap Task on Map TaskFollow11.98ScoreMEMPR8.36089.300410.2411.1796May 25, 2026Evaluation ResultsMethodMethodLinksScoreMEMPRModel=Qwen3 30B ThinkModel=Qwen3 30B Think2026.0511.98HumanModel=HumanModel=Human2026.0510.98COMPACTORModel=Llama 3 8BModel=Llama 3 8B2026.058.5