ResearchBenchmarksEmbodied Task Planning on WorldLines Downstream Embodied (planning set)For AI assistantsMarkdownContent indexEmbodied Task Planning on WorldLines Downstream Embodied (planning set)Follow68.4Plan Judge ScoreObsMem32.31241.68151.0560.419Jun 17, 2026Evaluation resultsMethodMethodLinksPlan Judge ScoreState ConsistencyPrecondition ValidityMemory UseObsMem2026.0668.472.170.269A-mem2026.0654.256.658.152.4Mem02026.0652.655.156.351.2GraphMem2026.0648.149.350.746.2MemoryOS2026.0633.736.137.631.9