ResearchBenchmarksReinforcement Learning on Ant v5 (Episodes to Threshold 2500)Reinforcement Learning on Ant v5 (Episodes to Threshold 2500)Follow890Episodes to Threshold 2500SAC872.8988.91,1051,221.1Mar 12, 2026Evaluation resultsMethodMethodLinksEpisodes to Threshold 2500SACConfiguration=H-EARSConfiguration=H-EARS2026.03890TD3Configuration=H-EARSConfiguration=H-EARS2026.03980SACConfiguration=VanillaConfiguration=Vanilla2026.031,240TD3Configuration=VanillaConfiguration=Vanilla2026.031,320