List your company

Conversation Labeling on Cross-agent 80 labeled sessions (val)

0.93Overall KappaCodex (OpenAI)
0.88840.89920.910.9208Mar 28, 2026

Evaluation results

MethodLinks
Codex (OpenAI)Best Ver.=v72026.030.93880.840.850.83
GeminiBest Ver.=v92026.030.93910.850.810.87
Claude CodeBest Ver.=v52026.030.91790.730.780.88
Human-HumanBest Ver.=–2026.030.89—0.780.730.7