Agent evaluation in MLflow (formerly Agent Evaluation)

Scores the quality of agents and GenAI apps with LLM judges, custom scorers and evaluation datasets in MLflow, and attaches the results to traces. It went GA as Mosaic AI Agent Evaluation and now lives in MLflow 3; automatic scoring of production traffic is a separate Beta feature.

Status and dates are extracted from public sources only. We don't see the private AE-shared roadmap. If something looks wrong, tell us.