ResearchBenchmarksWorld Modeling on ResourceAllocFollow17HSRAgent4.1047.45210.814.148Jun 26, 2026Evaluation ResultsMethodMethodLinksHSRScore Rate (SR)Tokens per TaskAgentBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.06171002,460GILPBackbone=GPT-4o-mini,...Backbone=GPT-4o-mini, n=20/202026.064.61003,126