ResearchBenchmarksGoal-oriented Robotic Manipulation on FetchPush 4-p v2Follow98.9Normalized ReturnTD3BC73.73280.26686.893.334Feb 11, 2026Evaluation ResultsMethodMethodLinksNormalized ReturnTD3BC2026.0298.9Flex-f-Q2026.0292.4IQL2026.0291.6Flex-f-DICE2026.0288.3Opti-DICE2026.0281.9CQL2026.0274.7