List your company

Reward Modeling on OSWorld Verified Class-Imbalanced Test Scripts 1.0 (test)

61.9PrecisionVAGEN
38.29244.42150.5556.679Jan 31, 2026

Evaluation results

MethodLinks
VAGENVerified Agent Backbone=UI-TARS-1.5-7B2026.0161.986.472.285
ZeroGUIVerified Agent Backbone=UI-TARS-1.5-7B2026.0152.371.660.478.9
FullTrajEvalVerified Agent Backbone=UI-TARS-1.5-7B2026.0150.569.158.377.8
AndroidGenVerified Agent Backbone=UI-TARS-1.5-7B2026.0145.348.146.775.3
WebRLVerified Agent Backbone=UI-TARS-1.5-7B2026.0144.2424375.1
DistRLVerified Agent Backbone=UI-TARS-1.5-7B2026.0143.440.74274.8
DigiRLVerified Agent Backbone=UI-TARS-1.5-7B2026.0139.238.338.772.9