ResearchBenchmarksPlanning Task on Star graph Stareasy synthetic (test)Follow100Sequence-level Exact Match AccuracyMDM29.59247.87166.1584.429Jun 8, 2026Evaluation ResultsMethodMethodLinksSequence-level Exact Match AccuracyMDM2026.06100ILM2026.06100DILM-S2026.06100DILM-M2026.06100ARM2026.0632.3