ResearchBenchmarksCredit Card Pick on Simulation EnvironmentsCredit Card Pick on Simulation EnvironmentsFollow85.4Success RateSCORE (DSRL)1.47223.26145.0566.839Jun 25, 2026Evaluation resultsMethodMethodLinksSuccess RateSCORE (DSRL)DSRL=trueDSRL=true2026.0685.4SCORE2026.0682.2Res-RL2026.0678.7FPO2026.0668.1RialToPost-distillation=truePost-distillation=true2026.0645Base2026.064.7