List your company

Reward Prediction on OSWorld Chrome

93.5Reward AccuracyPRORE
82.16485.10788.0590.993Sep 26, 2025

Evaluation results

MethodLinks
PROREEvaluator Agent=Claude-Sonnet-4.52025.0993.5
WebRLEvaluator Agent=Claude-Sonnet-4.52025.0987
StepCriticEvaluator Agent=Claude-Sonnet-4.52025.0987
DigiRLEvaluator Agent=Claude-Sonnet-4.52025.0984.8
DistRLEvaluator Agent=Claude-Sonnet-4.52025.0982.6