MLflow Project (LF Projects) / breaking change

MLflow 3 changes model storage and removes deprecated APIs

MLflow 3 promotes logged models to first-class entities, changes model logging from artifact_path to name, and stores model artifacts outside the run-artifact hierarchy. It also removes Recipes, several model flavors and parameters, the deployment-server command, and portions of the older evaluation API.

Affected product / version: MLflow 3 model storage and APIs
Announced: 2025-06-09
Effective: 2025-06-10
Source checked: 2026-09-07

Checked against the linked documentation on 2026-09-07. Recheck the vendor documentation before acting.

What to check in your application

PlaintextML interpretation of the documented change:

Pipelines and dashboards that discover models through run artifacts or list_artifacts need to move to Logged Model APIs and verify that lineage, retention, and access-control logic still reaches the intended artifacts.
Back up and migrate the database, upgrade the server, validate existing clients, then upgrade SDKs. Rehearse restoration: schema migrations may not be transactional, and newer clients may not work against older servers.

Limits and uncertainty

Dates differ: the launch post is June 9, PyPI availability June 10 (used as effective_on), and release pages June 11. The current guide includes later 3.x changes; verify your exact version. Client/server compatibility is best effort.

Official sources

Revision history

2026-09-07 — Initial record added after checking the linked official documentation; migration checks are editorial interpretation.

Found a problem? Keep the record URL, the claim that needs changing, and the supporting primary source together when reporting it to the publisher.