ResearchBenchmarksPolicy Optimization on GlucosePolicy Optimization on GlucoseFollow6.3True OutcomeORPO*5.9886.0696.156.231Apr 13, 2026Evaluation resultsMethodMethodLinksTrue OutcomeProxy PerformanceWorst-Case ResultORPO*Environment=GlucoseEnvironment=Glucose2026.046.3116.36-8.79Max-MinEnvironment=GlucoseEnvironment=Glucose2026.046.3102.66-1.71ORPO*2026.046.3116.36-8.79Max-Min2026.046.3102.66-1.71ORPOEnvironment=GlucoseEnvironment=Glucose2026.046100.48-27.54ORPO2026.046100.48-27.54