ResearchBenchmarksWorld Modeling on ToolChainFollow19.3HSRAgent1.7246.28710.8515.413Jun 26, 2026Evaluation ResultsMethodMethodLinksHSRSRTokens per TaskAgentBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.0619.31002,054GILPBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.062.41002,624