List your company

Agent Task Completion and Safety on AgentDojo Workspace stock (test)

81.4Utility ScoreCaMeL
49.305657.637865.9774.3022May 27, 2026

Evaluation results

MethodLinks
CaMeLModel=o4-mini-high2026.0581.40
LACUNAModel=gemini-2.5-pro2026.0556.254
LACUNAModel=o4-mini-high2026.0555.540
CaMeLModel=gemini-2.5-pro2026.0553.80
TACITModel=o4-mini-high2026.0552.860
TACITModel=gemini-2.5-pro2026.0550.540