Databricks releases, every cloud
- Level of detail expressions for custom calculations
Control aggregation granularity independently of visualization dimensions. Used for calculating percentages or comparing values to dataset-wide aggregates.
- Value sampling features renamed
Replaces value sampling, example values, and value dictionaries with prompt matching, format assistance, and entity matching.
- Dashboard navigation, pivot tables, and text widget improvements
Thicker scroll bar, larger title character limit, configurable pivot table header height, and markdown table support in text widgets.
- Databricks Runtime 18: January 15, 2026
It uses JDK 21 as the default Java Development Kit, replacing JDK 17.
- Databricks Runtime 18.0 is now GA
Replaces Databricks Runtime 7.3.
- Genie Agent (formerly Genie Space) and Databricks Assistant on Databricks on AWS GovCloud GA
Genie Agent and Databricks Assistant are now available on Databricks on AWS GovCloud.
- Run the change, do not just read it. Hands-on labs in your own Databricks workspace, graded when you submit. Browse labs
- Genie Agent (formerly Genie Space) sample data, Research Agent, and space manager improvements
Adds text answer summaries to API and increases example SQL title character limit to 1024.
- Genie One favorite and filtering improvements
Adds favorite and certified filters to listing pages, marked by a star icon.
- Manage Lakebase from the Lakebase App
Replaces Compute tab navigation in Lakehouse UI, accessed via apps switcher.
- Ingest Salesforce formula fields incrementally (Beta)
Replaces snapshotting with incremental ingestion, improving performance and reducing costs.
- Knowledge Assistant is now generally available
Available in select US regions for workspaces without Enhanced Security.
- Row filtering for managed ingestion connectors (Beta)
Applies conditions like a SQL WHERE clause, available for Google Analytics, Salesforce, and ServiceNow connectors.
- OpenAI GPT-5.1 Codex Max and Codex Mini now available as Databricks-hosted models
Access via Foundation Model APIs pay-per-token, subject to OpenAI's Acceptable Use Policy.
- OpenAI GPT-5.1 Codex Max and Codex Mini now available as Databricks-hosted models
Supports code generation and refactoring via Foundation Model APIs. Requires pay-per-token access.
- Databricks Runtime maintenance updates (01/09)
Includes bug fixes, security patches, and performance improvements for versions 12.2 to 17.3.
- Dashboard pivot tables, filters, and data export improvements
Pivot tables display total rows at top, Excel downloads in pivoted format.
- Genie Agent (formerly Genie Space) PDF export, citations, and accuracy improvements
Adds PDF export, SQL example citations, and improved currency formatting to Research Agent reports.
- Simplified serverless network security with Azure Network Security Perimeter
Uses AzureDatabricksServerless service tag, replacing subnet-based rules.
- Get this table by email. One Monday mail covering the week, filtered to the clouds and products you run. Weekly digest
- Automatic email notifications for expiring personal access tokens (GA)
Sends emails 14 days before token expiration, requires token lifetime to be set.
- Change default workspace access to Consumer access (Public Preview)
Preserves existing users' authoring permissions, streamlines consumer onboarding.
- Create skills for Databricks Assistant
Extends Databricks Assistant with domain-specific tasks, following the open Agent Skills standard.
- Databricks Assistant on the documentation site
Replaces search bar, requires no setup, powered by Genie.
- Databricks Assistant Agent mode is now in Public Preview
Automates multiple steps from a single prompt, using Azure OpenAI or Anthropic.
- Single-use refresh tokens for OAuth applications
Requires token rotation after each use, enhancing user-to-machine authentication flows.
- Anthropic Claude Haiku 4.5 now available as a Databricks-hosted model
Accessible via Foundation Model APIs on a pay-per-token basis.
- New Azure Databricks accounts will not have access to legacy features
Accounts created after December 18, 2025 use Unity Catalog exclusively.
- New Databricks accounts will not have access to legacy features
New accounts use Unity Catalog, lack DBFS root access. Applies to accounts created after December 18, 2025.
- Update request parameters for Delta Sharing recipient audit log events
Includes catalog_name, in addition to share_name.
- Connect to Lakebase from the SQL editor with read-write access
Supports full read-write access, see documentation for setup.
- Dashboard enhancements and fixes
Boolean values support custom string aliases, and filter widgets have several fixes and new features.
- Genie enhancements and fixes
Numeric columns are now formatted to handle IDs and currency. Entity matching scans up to 100 million rows.
- Lakebase metrics dashboard
Monitors system and database health with metrics like RAM and CPU usage.
- Meta Ads connector (Beta)
Ingests data from Meta Ads, requires setup as a data source.
- MySQL connector in Lakeflow Connect (Public Preview)
Enables incremental data ingestion from various MySQL databases. Requires access request through account team.
- Updated Lakeflow Jobs UI is generally available
Simplifies job creation and editing, see Lakeflow Jobs for details.
- View latest scheduled notebook job results
Shows latest scheduled run in notebooks and dashboards, can update notebook with latest results.
- Context based ingress control is now in Public Preview
Enables account admins to set allow and deny rules based on identity, request type, and network source.
- Flexible node types are now generally available
Falls back to alternative instance types when specified type is unavailable, improving launch reliability.
- Gemini 3 Flash now available as a Databricks-hosted model
Offers speed and scale for video analysis and data extraction. Requires no setup.
- Lakebase ACL support
Grants CAN CREATE or CAN MANAGE permissions from project settings.
Headlines, dates and product areas are Databricks' own, and every item links to the note it came from. The one-line summaries are ours.