ResearchBenchmarksCredit Card Pick on Real-world 1.0 (test)Credit Card Pick on Real-world 1.0 (test)Follow80Success RateRes-RL-3.218.44061.6Jun 25, 2026Evaluation resultsMethodMethodLinksSuccess RateRes-RL2026.0680SCORE (DSRL)variant=DSRLvariant=DSRL2026.0680SCORE2026.0680Base2026.0610RialTo2026.0610FPO2026.060