Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Important
This feature is in Beta. Account admins can control access to this feature from the account console Previews page. See Manage Azure Databricks previews.
Azure Databricks includes ready-to-use model services in Unity Catalog that serve a selection of high-quality generative AI and LLM foundation models. This article describes how to discover the model services and foundation models available to you, so you can start querying them without hosting your own models.
Discover model services
A model service is a Unity Catalog securable that represents a governed model endpoint. To get started, find a model service you can query. Two kinds are available:
- System-provided model services in the
systemcatalog andaischema (system.ai)—one for each Databricks-hosted foundation model, such assystem.ai.claude-opus-4-6. These are ready to query with no setup. - Custom model services that you or your teams create to route across multiple models or set up fallbacks. See Create custom model services.
You see the model services that you own or have the EXECUTE or MANAGE privilege on, with the required USE CATALOG and USE SCHEMA privileges. System-provided model services in system.ai are available to all account users by default; metastore admins can restrict access. To govern access, see Unity AI Gateway.
Unity AI Gateway
The Unity AI Gateway UI shows the model services you can access across catalogs, respecting catalog workspace bindings.
Catalog Explorer
Go to a schema to see its model services alongside your other Unity Catalog assets. The system.ai schema lists all system-provided model services.
REST API
List the model services in a schema with the Unity Catalog REST API. Pass the schema as the parent query parameter:
curl "https://<workspace-url>/api/2.1/unity-catalog/model-services?parent=schemas/<catalog>.<schema>" \
-H "Authorization: Bearer $DATABRICKS_TOKEN"
To list the system-provided model services, set parent=schemas/system.ai. The response returns a model_services array and a next_page_token that you pass back as page_token to retrieve the next page.
To get a single model service by its fully qualified name:
curl "https://<workspace-url>/api/2.1/unity-catalog/model-services/<catalog>.<schema>.<model-service>" \
-H "Authorization: Bearer $DATABRICKS_TOKEN"
Foundation models, availability, and terms
The system-provided model services in system.ai serve a selection of state-of-the-art foundation models that Azure Databricks hosts and pre-installs in Unity Catalog. The following apply equally to these foundation models and the model services that serve them:
- Region and feature availability: Foundation models and model services are available in regions enabled for Unity AI Gateway. See Model serving features availability.
- Model terms: Your use of these models and the services that serve them is subject to the applicable model terms and acceptable use policies. See Applicable model terms.