October 2026
Databricks released these features and improvements in October 2026.
Releases are staged. Your Databricks account might not be updated until a week or more after the initial release date.
DataFrame caching on serverless compute is now in Public Preview
October 23, 2026
DataFrame caching on serverless compute is now in Public Preview. You can cache a DataFrame with df.cache() or df.persist() on serverless compute to reuse an expensive result for the rest of your session. Caching requires serverless environment version 5 or above, Databricks Connect for Python 18.1.8 or above, or Databricks Connect for Scala 18.1.5 or above. The preview is enabled by default for most workspaces. Workspace admins can disable it from the Previews page. See Cache DataFrames on serverless compute.
Box Audit Logs managed ingestion connector (Beta)
October 23, 2026
The Box Audit Logs connector in Lakeflow Connect is now available in Beta. Use the connector to ingest enterprise audit events from Box into Databricks. See Box Audit Logs connector.
Screen MCP tool calls and results for jailbreaks (Beta)
October 21, 2026
You can now screen MCP tool calls and their results in Unity Gateway for jailbreak and prompt injection attempts. Apply the Jailbreak guardrail on input, output, or both. See Attach a built-in policy.
Lakeflow pipelines system tables are now generally available (GA)
October 21, 2026
The lakeflow system tables pipelines and pipeline_update_timeline are now generally available. These tables record the pipelines and pipeline updates in your account.
For a reference of these tables, see Jobs system table reference.
Genie Code tasks in Lakeflow Jobs are now generally available
October 21, 2026
Genie Code tasks in Lakeflow Jobs are now generally available. A Genie Code task runs your prompt autonomously as part of a job. It can read upstream task outputs and call tools as needed, and it returns a link to the resulting Genie Code chat with its response. See Genie Code task for jobs.
SurveyMonkey connector is available in Beta
October 20, 2026
The managed SurveyMonkey connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.
Use the connector to ingest surveys, survey structure, collectors, survey responses, and contacts from SurveyMonkey into Databricks. The connector uses OAuth user-to-machine (U2M) authentication and supports incremental ingestion, UI-based pipeline authoring, Declarative Automation Bundles, and API-based pipeline authoring.
Oracle Eloqua connector is available in Beta
October 20, 2026
The managed Oracle Eloqua connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.
Use the connector to ingest marketing data such as contacts, accounts, campaigns, email and form assets, form submissions, events, and website visitors from Oracle Eloqua into Databricks. The connector uses OAuth user-to-machine (U2M) authentication and supports incremental ingestion, UI-based pipeline authoring, Declarative Automation Bundles, and API-based pipeline authoring.
Asana connector is available in Beta
October 20, 2026
The managed Asana connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.
Use the connector to ingest work management data such as workspaces, teams, projects, tasks, portfolios, and goals from Asana into Databricks. The connector uses OAuth user-to-machine (U2M) authentication.
Saving an unchanged job no longer adds rows to the jobs system tables
October 19, 2026
Saving a job without changing its settings no longer adds rows to system.lakeflow.jobs or system.lakeflow.job_tasks, and change_time stays at the last change. This applies to saves from the UI, the Jobs API, and Declarative Automation Bundles. See Jobs table schema.
Filter data per end user with OAuth custom identity claims
October 19, 2026
You can now add a custom identity claim to a service principal's OAuth token and filter rows by it in a dynamic view or ABAC policy. Serve many end users through one service principal while each sees only their own data.
See Filter data per end user with OAuth custom identity claims.
DeepSeek V4.1 Flash is scheduled for retirement
October 19, 2026
DeepSeek V4.1 Flash will be retired on November 19, 2026. DeepSeek V4.1 Flash (1017) is the recommended replacement. See Deprecated and retired models for migration guidance.
DeepSeek V4.1 Flash (1017) is now available on Unity Gateway
October 19, 2026
DeepSeek V4.1 Flash (1017) is now available on Unity Gateway with text and image inputs. See supported regions. You can access this model using Foundation Model APIs.
Customers are responsible for ensuring their compliance with applicable terms.
Asynchronous processing with transformWithState is now generally available
October 19, 2026
Asynchronous processing for the Python row-based transformWithState API is now generally available. Run state operations and user logic concurrently for higher throughput with only minor code changes. See Asynchronous processing with transformWithState.
TypeSafe AI Jev 1.14.0 is now available on Unity Gateway
October 16, 2026
TypeSafe AI Jev 1.14.0 is now available on Unity Gateway. TypeSafe AI hosts it, and an account admin must enable the Externally-hosted TypeSafe AI Models preview. TypeSafe's terms govern its use.
READ METADATA privilege support for compliance security profile standards
October 16, 2026
The READ METADATA privilege now supports all compliance security profile standards supported by Databricks and is available by default for workspaces with the compliance security profile enabled.
For a full list of supported compliance standards, see Compliance security profile.
Delegate budget access with access control lists (ACLs)
October 16, 2026
Budgets now support access control lists (ACLs). Account and workspace admins can grant the Budget Manager or Budget Viewer role to users, groups, or Databricks service principals, delegating who can view and manage a budget. See Manage budget permissions.
ADBC driver for Power BI is now generally available
October 15, 2026
The Arrow Database Connectivity (ADBC) driver for the Databricks connector for Power BI is now generally available. ADBC is the default driver for new connections in Power BI Desktop and the Power BI service. See Configure ADBC or ODBC driver for Power BI.
MySQL integrated CDC pipeline is now generally available
October 15, 2026
The integrated CDC pipeline for MySQL is now generally available. An integrated CDC pipeline ingests change data from MySQL into Databricks using a single pipeline, combining extraction and application into one pipeline update. Unlike the standard gateway-based architecture, an integrated CDC pipeline requires no separate ingestion gateway. See Create an integrated CDC pipeline for MySQL.
Data Classification adds 50 new classes and support for 16 more countries and regions
October 15, 2026
Data Classification now detects 50 new classes of sensitive data, along with new regional support for 16 countries and regions:
- Payment card and financial data: Card security codes, card expiration dates, card track data, card PINs, SWIFT codes, and credit scores.
- Personal and employee data: Gender, residency status, disability data, sexual data, street addresses, geographic coordinates, compensation, and employment status.
- Software security secrets: Credentials and secrets used by software.
- Regional identifiers: Added 35 new classes and expanded coverage in existing classes across 18 countries and regions, including China, Mexico, the UK, other European countries, and more. See the documentation for details.
Regional tags are available in select regions. See Supported classification tags.
Consumer access to query Unity Gateway services is generally available
October 15, 2026
Users with the consumer access entitlement no longer need an account admin to enable the "Consumer access to Unity Gateway" preview to query the model services and model provider services shared with them, or to use Unity Gateway coding agent features. See Consumer access capabilities.
This change can increase model traffic from consumer access users, including traffic to external providers. To control this traffic, account admins can set budgets and rate limits, or disable direct model access, for consumer access users. See Manage budgets for Unity Gateway, Apply rate limits to model and MCP services, and Manage entitlements.
Redesigned MLflow Logs tab for AI Runtime runs
October 15, 2026
The MLflow run page Logs tab for AI Runtime runs now supports searching and filtering logs by node, level, and time range, following a running workload as new output arrives, and downloading logs. For new runs, it replaces the log files previously stored on the run's Artifacts tab, which are deprecated. Logs are available for 60 days. Logs for existing runs remain on the Artifacts tab. See View logs in the MLflow Logs tab.
The agentic code converter is generally available for SQL conversion
October 15, 2026
The agentic code converter is now generally available for the SQL target framework. Powered by Genie Code, it converts SQL from dialects such as T-SQL, Snowflake, Redshift, Oracle, BigQuery, and Teradata into Databricks SQL queries. You organize the work in a migration project, and the converter validates the output and fixes errors before writing the results to a workspace folder. After a conversion, you can continue chatting with Genie Code about the converted files to address action items or summarize the changes.
See Convert SQL with the agentic code converter.
Transactions support both managed Delta Lake and Apache Iceberg tables
October 14, 2026
Transactions are generally available for both Unity Catalog managed Delta Lake tables and Unity Catalog managed Apache Iceberg tables. Group multiple SQL statements across one or more managed Delta Lake or managed Apache Iceberg tables into a single atomic transaction using BEGIN ATOMIC ... END; or BEGIN TRANSACTION; ... COMMIT; syntax. Transactions require catalog commits enabled on participating tables.
See Transactions.
Data Classification API is now generally available
October 14, 2026
The Data Classification API and its Terraform resource are now generally available. Auto-tagging configuration (auto_tag_configs) remains in Public Preview. See the API reference.
Workspace-level package repositories for Databricks Apps are generally available
October 14, 2026
Workspace admins can now configure default Python and npm package repositories for Databricks Apps. Apps inherit the workspace configuration while still allowing app-level overrides.
See Configure default package repositories.
On-demand state repartitioning is now generally available
October 13, 2026
On-demand state repartitioning is now generally available. You can resize the number of shuffle partitions for a stateful Structured Streaming query or Lakeflow pipeline without losing checkpoint state. See On-demand state repartitioning for stateful streaming queries.
External writes to managed Delta tables are now generally available
October 13, 2026
Support for creating and writing to Unity Catalog managed Delta tables from external Apache Spark clients is now generally available. Tables must have catalog commits enabled. See Access Databricks tables from Delta clients.
Google Gemini 4 Argon is now available on Unity Gateway
October 13, 2026
Google Gemini 4 Argon is now available on Unity Gateway. You can access this model using Foundation Model APIs.
Customers are responsible for ensuring their compliance with applicable terms.
External access to views through the Unity REST API is generally available
October 13, 2026
External engines can now query Unity Catalog views over Delta Lake tables through the Unity REST API. See External access to views.
Cribl connector is available in Beta
October 13, 2026
The managed Cribl connector is now available in Beta in Databricks Lakeflow Connect.
Use the connector to ingest Cribl.Cloud Worker Group configuration, worker and health status, organization metadata, and Leader audit and access logs into Databricks. The connector uses OAuth machine-to-machine authentication and supports incremental ingestion for audit and access logs.
Shared output in clean rooms is now generally available
October 13, 2026
Clean room runs can now write to a shared output schema that every collaborator can read as soon as each table is created, even before the run finishes. Use it to share results, logs, and debugging information with all collaborators. See Share output with all collaborators.
JAR workloads in clean rooms are now generally available
October 13, 2026
You can now run compiled Java or Scala code in clean rooms. Run a JAR as its own clean room analysis, or register a function from a JAR and call it from a clean room notebook as a temporary UDF. See Run JAR workloads in Databricks clean rooms.
Create clean rooms with identity resolution providers in one step (GA)
October 13, 2026
You can now create a ready-to-use packaged clean room with an approved identity resolution provider, such as Acxiom, Adstra, LiveRamp, or The Trade Desk, in one step from the Clean Rooms page. The provider is preconfigured as the package provider, so you can add your data and run the provider's analysis right away. See Create a clean room with an identity resolution provider.
Databricks AI Runtime SDK replaces serverless_gpu
October 13, 2026
Starting in GPU environment version 6, the Databricks AI Runtime SDK (databricks-sdk-air) replaces the deprecated serverless_gpu package. It provides the @distributed and ray_init APIs and the databricks.air.data utilities under the databricks.air namespace. The gpus and gpu_type arguments are renamed to num_accelerators and accelerator_type. The old names still work in databricks.air but emit a FutureWarning, and support will end in version 7.
Serverless GPU environment version 6 includes both packages. Version 6 is the last version that includes serverless_gpu. Versions 5 and below continue to use it. See Migrate from serverless_gpu.
Unity Gateway Skills API and developer tools are available in Beta
October 12, 2026
The Unity Gateway Skills REST API is now available in Beta, with create, read, update, list, delete, and finalize support across Databricks developer tools:
- Databricks CLI
- Python SDK
- Go SDK
- Go Modular SDK
- Java SDK
- JavaScript SDK
See the Skills API and SDK reference.
Governed tag propagation is in Beta
October 12, 2026
Tag propagation in Unity Catalog is in Beta. When you enable tag propagation for a governed tag, Unity Catalog uses lineage to automatically apply that tag to derived tables and views whenever a supported operation runs, such as CREATE TABLE AS SELECT, CREATE VIEW, INSERT INTO ... SELECT, or MERGE. This keeps classification and ABAC enforcement consistent at both the table and column levels as data moves downstream.
See Tag propagation.
Parameters for Lakeflow pipelines are now generally available
October 12, 2026
Pipeline parameters are now generally available. Parameterize Lakeflow pipeline SQL source code with key-value pairs so a single pipeline can run against different catalogs, schemas, or tables without editing the source. Define parameter defaults in pipeline settings, override them when you start an update, and reference them in SQL with named parameter syntax. See Use parameters with pipelines.
Private connectivity from serverless compute to AWS-managed resources is now generally available
October 12, 2026
AWS Private Link from Databricks serverless compute to AWS-managed resources, such as S3 buckets, previously in Public Preview, is now generally available. Your data traffic remains entirely within the AWS network and never traverses the public internet. See Configure private connectivity to AWS-managed resources.
Agent memory pipelines are in Beta
October 12, 2026
Agent memory pipelines are in Beta. A memory pipeline connects a session store to a memory store. Each run uses an LLM to distill durable facts, preferences, and decisions from stored conversations into memory entries that your agent recalls in later conversations. You can run a pipeline on demand or about every 24 hours. See Agent memory pipelines.
Knowledge Assistant and Supervisor Agent are now legacy
October 12, 2026
Knowledge Assistant and Supervisor Agent are now legacy offerings. For new agents written in code, Databricks recommends the Agent Bricks CLI. To build agents without code, use Genie Agents. See Legacy agent offerings.
Feature Views are now generally available
October 12, 2026
Feature Views are now generally available for batch, streaming, and request-time data. They will be enabled for compliance security profile workspaces in mid-November 2026. See Feature Views.
Context-based ingress policies for inbound private connectivity are now generally available
October 12, 2026
Context-based ingress policies for inbound Private Link are now generally available. Control which registered VPC endpoints can reach your workspaces, and combine endpoint rules with identity and request type conditions in the same policy. If Context-Based Ingress: Workspace Private Access Policies isn't already enabled for your account, an account admin can enable it from the account console Previews page. See Configure private access policy.
Use request tags in Unity Gateway service policies (Beta)
October 9, 2026
Custom service policies in Unity Gateway can now check caller-supplied request tags, such as requiring a tag before a model request or MCP tool call proceeds. See Request tags.
Use OpenJev as the evaluator for LLM-as-a-judge service policies (Beta)
October 9, 2026
You can now select OpenJev (Qwen3.5 4B) as the evaluator model service of a custom LLM-as-a-judge service policy. OpenJev classifies content without generating output tokens, so it typically responds faster than a chat evaluator. See Use OpenJev as the evaluator.
Google Workspace connector is available in Beta
October 9, 2026
The managed Google Workspace connector is now available in Beta in Databricks Lakeflow Connect.
Use the connector to ingest audit activity from Google Workspace applications and services into Databricks. The connector uses OAuth user authorization and supports incremental ingestion.
Unity Gateway supports Amazon Bedrock Mantle
October 9, 2026
You can now access Amazon Bedrock Mantle models through Unity Gateway model provider services, with Unity Catalog permissions, rate limits, budgets, and inference tables. See Model providers external to Databricks.
Productboard connector is available in Beta
October 8, 2026
The managed Productboard connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.
Use the connector to ingest product hierarchy, roadmap, objectives, customer feedback notes, and workspace member data from Productboard into Databricks. The connector uses Server-to-Server OAuth authentication and supports incremental ingestion for notes, UI-based pipeline authoring, Declarative Automation Bundles, and API-based pipeline authoring.
Microsoft Entra ID connector is available in Beta
October 22, 2026
The managed Microsoft Entra ID connector is now available in Beta in Databricks Lakeflow Connect.
Use the connector to ingest Microsoft Entra ID users, groups, applications, service principals, directory roles, identity configuration, and audit and sign-in logs into Databricks. The connector uses OAuth machine-to-machine authentication and supports incremental ingestion for audit and sign-in logs.
The Discover page and domains are now GA
October 8, 2026
The Discover page and domains are now generally available. Discover is a curated browsing experience that helps users find assets across Databricks without requiring deep knowledge of catalog hierarchies. Use domains and subdomains to organize assets such as tables, dashboards, notebooks, and Genie Agents by business area, highlight key assets with custom curation, and explore AI-recommended assets.
Azure DevOps connector is available in Beta
October 8, 2026
The managed Azure DevOps connector is now available in Beta in Databricks Lakeflow Connect.
Use the connector to ingest project, repository, work item, build, release, and identity data from Azure DevOps into Databricks. The connector authenticates with a Microsoft Entra service principal and supports incremental ingestion of commits.
Enforce partner guardrails with external service policies (Beta)
October 7, 2026
You can now enforce a third-party guardrail on Unity Gateway traffic. An external service policy sends request and response content to your guardrail vendor and blocks the call when the vendor denies it, with no changes to your applications. See Enforce a partner guardrail with an external service policy.
DeepSeek V4.1 Flash supports priority pay-per-token and reserved provisioned throughput
October 7, 2026
You can now use DeepSeek V4.1 Flash with priority pay-per-token for latency-sensitive traffic or reserved provisioned throughput for dedicated capacity.
Dataset materialization for AI/BI dashboards is now in Beta
October 7, 2026
You can now pre-compute and store dashboard dataset results on a schedule to make dashboard interactivity much faster. Published dashboards render from stored results instead of rerunning expensive queries on every load or interaction. See Dataset materialization in AI/BI dashboards.
Anthropic Claude Haiku 5.5 is now available on Unity Gateway
October 7, 2026
Anthropic Claude Haiku 5.5 is now available on Unity Gateway. You can access this model using Foundation Model APIs.
Customers are responsible for ensuring their compliance with applicable terms.
Metric view window measures are generally available
October 6, 2026
Window measures in metric views are now generally available. Window measures define rolling window, cumulative, or semiadditive aggregations, such as moving averages, period-over-period changes, and running totals. See Window measures.
Zhipu AI GLM 5.2 is now a legacy model
October 6, 2026
GLM 5.2 is now a legacy model. Databricks recommends migrating to GLM 5.3 or GLM 5.3 Flash.
Genie Code CLI is in Beta
October 6, 2026
Genie Code CLI is a coding agent that runs in your terminal, tuned for data and AI work on Databricks. It works with your local files and uses the Databricks CLI to discover data, answer questions, and build and deploy assets such as pipelines, models, and apps. Model access is provided and governed through Unity Gateway. See Genie Code CLI.
DeepSeek V4 Flash (0731) is scheduled for retirement
October 6, 2026
DeepSeek V4 Flash (0731) will be retired on November 5, 2026. DeepSeek V4.1 Flash is the recommended replacement. See Deprecated and retired models for migration guidance.
Use multiple triggers for Lakeflow Jobs (Beta)
October 5, 2026
Lakeflow Jobs can now use up to five triggers. You can combine trigger types, add multiple triggers of the same type, and pause each trigger independently. See Add multiple triggers to a job.
Run Lakeflow pipelines as a group
October 5, 2026
You can now set a pipeline's run-as identity to an account group, in addition to a user or a service principal. Pipeline updates run with the group's permissions, so a team can own an automated pipeline without tying it to one person's account. See Set the run-as identity to a group.
Genie Code scheduled tasks are now available by default for workspaces with the compliance security profile enabled
October 5, 2026
Genie Code scheduled tasks are now available by default for workspaces with the compliance security profile enabled.
See Genie Code scheduled tasks.
ai_parse_document now extracts chart data as JSON
October 5, 2026
ai_parse_document now extracts chart values as structured JSON and describes their series, legends, and axes. Search and agents can use these values in answers. See ai_parse_document function.
Agent Bricks documentation for code-first agent development
October 5, 2026
The Agent Bricks documentation now centers on building, deploying, and running custom agents in code. See Develop agents on Databricks:
- Agent compute stack: New pages describe how a deployed agent's framework or harness, agent server, and agent runtime fit together. See Deploy agents on Databricks.
DurableAgentServerand Agent Runtime:DurableAgentServerserves agents through the invocation API, with synchronous, streaming, and background runs, persistent run state, and crash recovery. Agent Runtime runs it on Databricks Apps. See Agent Server and Agent Runtime.- Querying agents: The query guide is organized by agent server, starting with the invocation API. See Query agents deployed on Databricks.
- Legacy offerings: Offerings that are no longer recommended for new agents, including the MLflow
AgentServer,LongRunningAgentServer, and agents deployed to Model Serving, are now in a Legacy section that maps each one to its replacement. See Legacy agent offerings.
OpenJev (Qwen3.5 4B) is now available on Unity Gateway
October 2, 2026
OpenJev (Qwen3.5 4B) is available as a pay-per-token model service. Use the TypeSafe System One API to evaluate yes or no, choice, and scoring questions and get structured answers and probabilities. See model availability by region.
Customers are responsible for ensuring their compliance with applicable model terms.
JDBC Unity Catalog connections are now generally available
October 2, 2026
JDBC Unity Catalog connections are now generally available on serverless compute, Databricks Runtime 19 and above, and SQL warehouses 2025.40 and above, including support for SSL certificates. OAuth M2M authentication remains in Beta. See JDBC connection.
Save reusable query templates in Google Sheets
October 2, 2026
The Databricks Connector for Google Sheets now lets you save imports as a reusable template that you and your team can run in one step, each with your own Unity Catalog permissions. See Templates in Google Sheets and Query Databricks data in Google Sheets.
Google Gemini 2.5 Pro is retired
October 2, 2026
Google Gemini 2.5 Pro is no longer available. The recommended replacement models are Gemini 3.1 Pro or Gemini 3.5 Flash. See Deprecated and retired models for migration guidance.
Google Gemini 2.5 Flash is retired
October 2, 2026
Google Gemini 2.5 Flash is no longer available. The recommended replacement models are Gemini 3.1 Pro or Gemini 3.5 Flash. See Deprecated and retired models for migration guidance.
Customer-managed keys for query history are now generally available
October 2, 2026
You can now use a customer-managed key on the system catalog to decrypt statement_text and error_message in the query history system table. See Reading encrypted fields.
ai_parse_document now extracts chart data
October 2, 2026
ai_parse_document now returns more detailed figure descriptions. For charts, the description also includes the chart data as structured JSON, such as series values and legend labels. See ai_parse_document function.
Standard ai_extract now delivers higher-quality extraction
October 2, 2026
The standard (default) mode for ai_extract now uses a new model checkpoint that is 2.3x faster while improving overall quality at no additional cost. See the ai_extract documentation.
GitLab connector is available in Beta
October 1, 2026
The managed GitLab connector is now available in Lakeflow Connect in Beta. Workspace admins can control access to this feature from the Previews page.
Use the connector to ingest repository, project, group, CI/CD, and security data from GitLab into Databricks. The connector uses OAuth U2M authentication and supports incremental ingestion, UI-based pipeline authoring, Declarative Automation Bundles, and API-based pipeline authoring.
Schedule automatic data refresh in the Databricks Excel Add-in (Public Preview)
October 1, 2026
You can now schedule automatic refreshes of Databricks data imported into Microsoft Excel. Choose an interval or calendar schedule and the SQL warehouse that runs each refresh, then review recent runs from the Databricks Excel Add-in. See Schedule automatic data refresh in the Databricks Excel Add-in.
Cross-engine attribute-based access controls (ABAC) is now generally available
October 1, 2026
Cross-engine attribute-based access controls (ABAC) is now generally available. Supported external engines can read Unity Catalog managed tables with ABAC policies, row filters, and column masks enforced by Databricks serverless compute.
See Cross-engine attribute-based access controls (ABAC) for requirements, supported clients, and serverless compute costs.