Product watch
Azure Databricks release notes · September 2026
View all azure databricks release notes
September 2026#
- Git Folder Serverless is in Beta — Sep 02 · Azure · Product You can now use Git Folder Serverless to share serverless compute and a pyproject.toml -managed environment across notebooks and files in a Git folder. See Git Folder Serverless.
- Google Gemini 3.8 Flash now available as a Databricks-hosted model — Sep 02 · Azure · Product Model Serving now supports Google Gemini 3.8 Flash as a Databricks-hosted model. You can access this model using Foundation Model APIs. Customers are responsible for ensuring their compliance with applicable terms.
- AI Search supports struct and map columns and filter-only queries — Sep 02 · Azure · Product You can now index and filter struct and map columns in AI Search, and run filter-only queries that return rows matching a filter without vector or keyword retrieval. See Query an AI Search index.
- Apache Arrow support for Zerobus Ingest is now generally available — Sep 01 · Azure · Product Apache Arrow support for Zerobus Ingest in Lakeflow Connect is now generally available. See Use Arrow Flight with Zerobus Ingest.
- HITRUST and IRAP compliance controls are now generally available — Sep 01 · Azure · Product HITRUST and IRAP compliance controls are now generally available through the compliance security profile. In addition, HIPAA, HITRUST, and IRAP now require the compliance security profile. See HITRUST and Infosec Registered Assessors Program (IRAP).
- Databricks Runtime maintenance updates (09/01) — Sep 01 · Azure · Product New maintenance updates are available for supported Databricks Runtime versions. These updates include bug fixes, security patches, and performance improvements. For details, see: Databricks Runtime 18 LTS Databricks Runtime 18.2 Databricks Runtime 18.1 Databricks Runtime 17.3 LT
- Anthropic Claude Fable 5.1 now available as a Databricks-hosted model — Sep 01 · Azure · Product Unity AI Gateway now supports Anthropic Claude Fable 5.1 as a Databricks-hosted model. You can access this model using Foundation Model APIs. Customers are responsible for ensuring their compliance with applicable terms.