ResearchTasksConversational EvaluationConversational EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAlpacaEval 2.0CPO26.57LC14May 21, 2026CoReflect Simulation Iteration t=1 1.0Gemini 2.5 Pro4.79ODI (Output Delivery Integrity)7Feb 26, 2026CoReflect 1.0 (Iteration 3)Gemini 2.5 Pro4.86ODI7Feb 26, 2026CoReflect Iteration 2 (test)Gemini 2.5 Pro4.88ODI7Feb 26, 2026Arena-Hard 0.1VAR10.8WR (%)3Mar 24, 2026