In the MLflow Model Registry, which sequence correctly represents the typical model lifecycle stages?
Select an answer to reveal the explanation.
Short Explanation and Infographic
The registry lifecycle is like a job application — you start with no status (None), move through a trial period (Staging), land the role (Production), and eventually get filed away (Archived).
Full explanation below image
Full Explanation
The MLflow Model Registry defines four official stages: None (newly registered model version, not yet assigned a stage), Staging (model is being tested and validated, suitable for QA or pre-production environments), Production (model is live and serving real traffic), and Archived (model is no longer active but preserved for audit/rollback purposes). These stages are managed via the UI or API calls like MlflowClient().transition_model_version_stage(). Models are not deleted from the registry when archived — they remain queryable. The other options use made-up stage names not part of the MLflow spec.