Databricks Release Notes

Route Claude Code through Unity Gateway using an existing Claude Pro, Max, Enterprise, or Team subscription, at no additional model cost, while your organization keeps centralized governance, usage tracking, and access control. Databricks relays your Anthropic OAuth token on each request, so no credentials are stored, and requests bill against your Anthropic subscription. Access is governed by standard Unity Catalog privileges on the model provider you register, and gateway features such as usage tracking, inference tables, rate limits, and spend tracking keep working.

See Claude Max and Anthropic Enterprise support.

Genie Code is now available on workspaces that have the compliance security profile enabled with the HIPAA compliance standard selected and partner-powered AI features disabled. On these workspaces, Genie Code is served by Databricks-hosted models rather than partner-powered models.

For the full list of AI assistive features affected by the partner-powered AI features setting and the hosting provider used by each, see Partner-powered AI features.

Automatic upgrades begin rolling out row tracking and Checkpoint V2 to existing Unity Catalog managed tables, not just new tables. This rollout is gradual, so these features will be available to different customers at different times.

Automatic upgrades turn on a feature only after the observation window verifies that every client accessing a table supports it. To track which features are available for automatic upgrades, see Supported features.

Registering Python user-defined table functions (UDTFs) in Unity Catalog is now generally available in Databricks Runtime 18. A Unity Catalog Python UDTF is a governed function that returns a table, invoked in the FROM clause of a SQL statement to return multiple rows and columns. Unity Catalog Python UDTFs run on serverless compute, classic compute with standard access mode, and SQL warehouses.

See Python user-defined table functions (UDTFs) in Unity Catalog and Databricks Runtime 18 LTS.

You can now write and run Python unit tests for your Spark Declarative Pipelines on Lakeflow transformation logic directly in the Lakeflow Pipelines Editor. Validate Python and SQL transformations against mock data using catalog-table redirection, including Auto CDC, streaming tables, expectations, and append flows. Test edge cases and iterate using mock inputs for supported table-identifier operations; review the isolation limitations before running tests. This feature is in Beta.

See Unit testing for pipelines.

Dedicated group clusters will be available by default for workspaces with the compliance security profile enabled in September 2026.

You can assign a compute resource to a group using the Dedicated access mode, letting group members share compute while securely running languages and workloads not supported by standard access mode, such as Databricks Runtime for ML, RDD APIs, and R. When a user connects to a group cluster, their permissions down-scope to the group's permissions for the cluster session.

See Dedicated compute group access.

Lakebase adds AI-assisted troubleshooting: database issues and early warnings are surfaced to you with Insights, and you can then investigate and fix them conversationally with Genie. Both draw on database telemetry captured to Unity Catalog. See Observability.

The system.tags.governed_tags system table is in Beta. This table contains a record for each governed tag key in your account, including customer-created and system-generated tags. Use it to programmatically determine which tags in your account are governed.

See Governed tags system table reference.

Git CLI commands in Git folders are now in Public Preview. Git folders that you create from the UI automatically get Git CLI access when your workspace is eligible, so you can run standard Git commands such as git stash, interactive rebase, pre-commit hooks, Git submodules, and Large File Storage (LFS) directly from a Databricks terminal. Workspace admins can control access to this feature from the Previews page. See Use Git CLI commands.

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 LTS
• Databricks Runtime 16.4 LTS
• Databricks Runtime 15.4 LTS
• Databricks Runtime 14.3 LTS
• Databricks Runtime 13.3 LTS

Databricks Model Serving now supports Anthropic Claude Opus 5 as a Databricks-hosted model.

To access this model, use:

• Foundation Model APIs pay-per-token
  • Query reasoning models
  • Query vision models

ai_parse_document reduces latency and improves the accuracy of layout and bounding box detection. See ai_parse_document function.

A C++ SDK for the Zerobus Ingest connector in Lakeflow Connect is now available in Beta. The SDK provides a high-throughput gRPC client for ingesting records directly into Unity Catalog Delta tables, with support for JSON and Protocol Buffers. See Use Zerobus Ingest.

You can now schedule Genie Code to run a prompt on a recurring schedule. Each run executes your prompt as a full Genie Code session and produces a continuable Genie chat with the results, so you can turn recurring analysis into an unattended background task. Create a scheduled task by asking Genie Code in chat or by filling in the create form manually. For more information, see Genie Code scheduled tasks.

Databricks Runtime 19 and Databricks Runtime 19 for Machine Learning are now generally available (GA), powered by Apache Spark 4.2.0.

For release notes, see Databricks Runtime 19 and Databricks Runtime 19 for Machine Learning.

With the general availability of Databricks Runtime 19, Databricks Runtime 18 and Databricks Runtime 18 for Machine Learning have transitioned to long-term support (LTS). LTS versions receive three years of stability and security fixes.

For release notes, see Databricks Runtime 18 LTS and Databricks Runtime 18 LTS for Machine Learning. For lifecycle details, see Databricks Runtime support lifecycles.

Variant will be available by default for workspaces with the compliance security profile enabled in July 2026.

Use Variant to ingest semi-structured data from streaming sources (Kinesis, Event Hub), REST APIs, and schemaless databases, while maintaining high-performance query access.

See Variant type support for Apache Iceberg and Delta Lake.