ResearchBenchmarksPlanning Task on Star graph task Starmedium synthetic (test)Follow100Exact Match Accuracy (Sequence-level)ILM33.9651.10568.2585.395Jun 8, 2026Evaluation ResultsMethodMethodLinksExact Match Accuracy (Sequence-level)ILM2026.06100DILM-S2026.06100DILM-M2026.0698.6ARM2026.0675MDM2026.0636.5