ResearchBenchmarksWorld Modeling on TaskGraphFollow14.2HSRAgent3.5926.3469.111.854Jun 26, 2026Evaluation ResultsMethodMethodLinksHSRSRTokens/TaskAgentBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.0614.21002,320GILPBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.0641002,904