Releasesby GlyphUpcoming removal of AI Functions support on older runtimes
Starting January 2027, AI Functions in SQL queries will require Databricks Runtime 17.3 LTS or higher. This change does not affect ai_similarity or the AI Functions REST API.
Agent Bricks and Mosaic AI Release notes from Databricks
88 releases · 7 this month
AI Functions require Databricks Runtime 17.3 LTS or above
Databricks announced that starting January 2027, AI Functions requiring Databricks Runtime 15.4 LTS or higher (excluding ai_similarity) will mandate Databricks Runtime 17.3 LTS or above.
Managed warehouse for MLflow (Beta)
Managed warehouse for MLflow is available in Beta, enabling users to view Unity Catalog traces in the MLflow UI without manually choosing a SQL warehouse.
AI Functions usage is now included in Unity Gateway budgets
Beginning November 1, 2026, Unity Gateway budgets include spend from task-specific AI Functions such as ai_parse_document , ai_extract , and ai_classify across batch and real-time execution paths. See Manage budgets for Unity Gateway .
Claude Code auto mode for Anthropic and non-Anthropic models
Claude Code auto mode is now available for both Anthropic and non-Anthropic models hosted on Databricks, reducing permission prompts for tool calls.
View personal and shared spend limits with the ug CLI
The ug CLI updated `ug usage` to display up to five active spend limits ordered by highest percentage used.
OpenAI GPT-6.1 Terra is now available on Unity Gateway
Databricks now supports the OpenAI GPT-6.1 Terra model via Foundation Model APIs on Unity Gateway.
Check budget usage with the Spend Status API
The new Spend Status API allows users to monitor usage against Unity Gateway budget limits. Developers can retrieve the status for individual budgets or list statuses across multiple budgets.
Use Claude desktop and ChatGPT desktop work agents with Unity Gateway
You can now use ug to configure the Claude desktop and ChatGPT desktop work agents with governed models, MCP servers, and skills. Subscription relay is available only in Claude desktop. Learn more .
Screen MCP tool calls and results for jailbreaks (Beta)
Databricks introduced a public beta feature in Unity Gateway allowing developers to screen MCP tool calls and results for jailbreak and prompt injection attempts using input/output policies.
ai_enrich can ground rows with external MCP search providers (Beta)
The ai_enrich function can ground each row with a search tool on an external Model Context Protocol (MCP) server that you register in Unity Gateway, such as You.com or Exa. Add an mcp knowledge source alone or alongside other knowledge sources in the same call. This feature is in Beta .
DeepSeek V4.1 Flash (1017) is now available on Unity Gateway
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 .
TypeSafe AI Jev 1.14.0 is now available on Unity Gateway
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.
Consumer access to query Unity Gateway services is generally available
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 .
Google Gemini 4 Argon is now available on Unity Gateway
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 .
Agent memory pipelines are in Beta
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 .
Use OpenJev as the evaluator for LLM-as-a-judge service policies (Beta)
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 .
Enforce partner guardrails with external service policies (Beta)
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 .
Agent Bricks documentation for code-first agent development
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 .
Google Gemini 2.5 Flash is retired
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.
Agent Bricks CLI is in Beta
The Agent Bricks CLI ( databricks-agentbricks ) is a command-line tool for building and deploying custom code agents on Databricks. Scaffold a project from a bundled framework template, run it locally, and deploy it to the Databricks agent runtime, with managed memory, sessions, tools, and MLflow tracing wired up from one authenticated command.
xAI Grok 4.7 is now available on Unity Gateway
xAI Grok 4.7 is now available on Unity Gateway. You can access this model using Foundation Model APIs pay-per-token . Customers are responsible for ensuring their compliance with applicable terms .
Usage tracking adds session metadata and token details (Public Preview)
The system.ai_gateway.usage system table now includes coding-agent session metadata and invocation source. New token-detail fields capture tokens written to the 5-minute and 1-hour prompt caches, along with file and web search counts. See Track model usage .
Anthropic Claude Opus 5.5 is now available on Unity Gateway
Anthropic Claude Opus 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 .
Back model provider service credentials with a Unity Catalog secret
Back a model provider service's API key with a customer-owned Unity Catalog secret. The service reads it at query time, so rotating the secret takes effect with no service change. See Back a provider credential with a Unity Catalog secret .
Configure provider discounts for external model spend
Unity Gateway estimates external model spend using each provider's published list prices. If your organization has negotiated different pricing with a provider, you can now configure a price multiplier on the model provider service so that spend estimates reflect those rates.
Unity Gateway support for compliance security profile standards
Unity Gateway 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 .
Managed agent memory and sessions are in Beta
Managed agent memory and managed agent sessions are in Beta . Databricks now provides two fully managed, Lakebase -backed stores for agent state, both usable from agents built on any framework: Managed agent memory gives agents durable, long-term memory that persists across conversations, with semantic search. See Managed agent memory .
DeepSeek V4.1 Flash is now available on Unity Gateway
DeepSeek V4.1 Flash 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 .
OpenAI GPT-6 Astra is now available on Unity Gateway
OpenAI GPT-6 Astra is now available on Unity Gateway for enterprise reasoning, structured document processing, coding, agentic search, multimodal inputs, and long-context workloads. You can access this model using Foundation Model APIs pay-per-token . Customers are responsible for ensuring their compliance with applicable terms .
Google Gemini 3.8 Flash now available as a Databricks-hosted model
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 .
Anthropic Claude Fable 5.1 now available as a Databricks-hosted model
Unity 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 .
Monitor and cap external provider costs
Unity Gateway can now track and cap spend on external model providers, such as Amazon Bedrock and Azure AI Foundry, not just Databricks-provided models. If your organization already has capacity commitments directly with a provider, Unity Gateway becomes the single point to track that usage and enforce spend limits across teams and use cases.
AI Functions REST API is now generally available
You can now call ai_parse_document , ai_extract , and ai_classify through the AI Functions REST API. See intelligent document processing for more details.
Zhipu AI GLM 5.3 Flash is now available on Unity Gateway
Zhipu AI GLM 5.3 Flash is now available on Unity Gateway for multimodal reasoning, coding, and agentic workloads. Access it through the same governed API as other models, with centralized permissions, guardrails, spend controls, and observability, and switch models without changing application code.
Unity Gateway now supports Provisioned Throughput
Govern Provisioned Throughput (PT) traffic through Unity Gateway, bringing PT models under the same access controls, usage tracking, rate limits, payload logging, and guardrails as other model traffic. See AI governance with Unity Gateway .
Consumer access to query Unity Gateway services will become generally available
Querying Unity Gateway model services and model provider services with Consumer access is currently opt-in through the Consumer access to Unity Gateway Public Preview . When this capability becomes generally available, consumer access users will be able to query these services without the preview enabled.
DeepSeek V4 Pro (0813) is now available on Unity Gateway
DeepSeek V4 Pro (0813) is now available on Unity Gateway for complex reasoning and agentic coding workflows. Access it through the same governed API as other models, with centralized permissions, guardrails, spend controls, and observability, and switch models without changing application code.
Smart Routing: Cut AI costs by routing each task to the right model
Smart Routing (Beta) automatically picks the right model and agent harness for each task, based on complexity, permissions, and budget. It shifts routine work to lower-cost models and reserves frontier models for hard problems, and every routing decision is visible and auditable.
Custom trace views in the MLflow trace explorer are in Beta
You can now build custom trace views in the Databricks MLflow trace explorer. Describe the layout you want in plain language, and Genie generates a reusable view that surfaces the trace fields, metrics, and feedback controls most relevant to your review workflow. See Build a custom trace view .
ABAC GRANT policies now cover Unity Gateway resource types (Beta)
Attribute-based access control (ABAC) GRANT policies dynamically grant Unity Catalog privileges to securables whose governed tags match a condition. They now support model services, model provider services, Model Context Protocol (MCP) services, and agent services, which are the resource types Unity Gateway governs.
DeepSeek V4 Flash (0731) is now available on Unity Gateway
DeepSeek V4 Flash (0731) is now available on Unity Gateway for fast, cost-efficient reasoning, coding, and agentic tool use. Access it through the same governed API as other models, with centralized permissions, guardrails, spend controls, and observability, and switch models without changing application code.
Unity Gateway API: Manage resources with SDKs, the CLI, and Terraform
You can manage Unity Gateway resources programmatically instead of clicking through the UI, with full create, read, update, list, and delete support for endpoints, model providers, and MCP servers.
Unity Gateway is now generally available
Unity Gateway is now generally available. Unity Gateway is the Databricks governance solution for enterprise AI, part of Unity Catalog. With Unity Gateway you can: Control which AI services teams can use. Route and manage AI traffic across providers. Govern MCP servers to control access and costs. Monitor usage, cost, access, and lineage from one place.
Use Unity Gateway with your own Claude enterprise subscription
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.
Upcoming general availability of Unity Gateway
Unity Gateway, the Databricks governance solution for enterprise AI, is scheduled for general availability soon. Built on Unity Catalog, Unity Gateway extends governance to the runtime interactions between models, agents, MCP servers, and tools.
Anthropic Claude Opus 5 now available as a Databricks-hosted model
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
Priority pay-per-token for Foundation Model APIs
Priority pay-per-token, also referred to as priority mode, for Foundation Model APIs pay-per-token is generally available for partner models and in preview for open source models. Priority mode is for latency-sensitive, real-time applications that need more consistent availability under load.
Google Gemini 3.5 Flash Lite now available as a Databricks-hosted model
Model Serving now supports Google Gemini 3.5 Flash Lite as a Databricks-hosted model. You can access this model using Foundation Model APIs . Customers are responsible for ensuring their compliance with applicable terms .
Connect and govern any external model provider
Connect and govern any external model provider (such as OpenAI, Anthropic, Amazon Bedrock, and newly released models like Meta's Muse Spark 1.1) through Unity Gateway. Register a provider once as a governed Unity Catalog securable, and every team can query it through Unity Gateway with full governance and observability.
Google Gemini 3 Pro is retired
Google Gemini 3 Pro is retired. Gemini 3.1 Pro is the recommended replacement. See Deprecated and retired models for migration guidance.
Databricks-provided MCP Services will GA soon
Databricks-provided MCP Services for SaaS applications such as Slack, GitHub, Atlassian, and Google Drive will soon be generally available, with both read and write operations enabled by default. Admins and users can disable MCP Services and individual tools in Unity Gateway . See Connect agents to tools with MCP Services .
Protobuf tensor input for custom model serving endpoints (Public Preview)
For custom model serving endpoints with a tensor-based input signature, you can now send the request body as a serialized KServe v2 ModelInferRequest with Content-Type: application/x-protobuf , instead of JSON. This feature is in Public Preview and is only available on endpoints deployed after July 9, 2026. See Protobuf tensor input .
Google Gemini 2.5 Flash and Pro are scheduled for retirement
Google Gemini 2.5 Flash and Gemini 2.5 Pro will be retired on October 2, 2026. The recommended replacement models are Gemini 3.1 Pro or Gemini 3.5 Flash. See Deprecated and retired models for migration guidance.
Google Gemini 3.1 Flash Image and Gemini 3 Pro Image now available as Databricks-hosted models
Model Serving now supports Google Gemini 3.1 Flash Image and Gemini 3 Pro Image as Databricks-hosted models. You can access these models using Foundation Model APIs . Customers are responsible for ensuring their compliance with applicable terms .
Anthropic Claude Sonnet 5 now available as a Databricks-hosted model
Model Serving now supports Anthropic Claude Sonnet 5 as a Databricks-hosted model. Designed for coding, agentic workflows, and professional work at scale, Claude Sonnet 5 delivers near-Opus-level intelligence with Sonnet's cost efficiency and speed. You can access this model using Foundation Model APIs .
About and activity
Agents, model serving, the AI Gateway, vector search and MLflow.
Other Databricks products: AI/BI and Genie, Databricks Apps, Databricks Data Intelligence Platform, Databricks SQL, Delta Lake, Lakebase, Lakeflow, MLflow, Unity Catalog
Releases per month
1Jun
17Jul
17Aug
19Sep
27Oct
7Nov
Official release-notes pages 14
| Release-notes page | Site | Feed |
|---|---|---|
| AI/BI (dashboards, Genie) docs.databricks.com | docs.databricks.com | None |
| Databricks CLI releases github.com | github.com | RSS |
| Databricks Connect docs.databricks.com | docs.databricks.com | None |
| Databricks Runtime docs.databricks.com | docs.databricks.com | None |
| Declarative Automation Bundles docs.databricks.com | docs.databricks.com | None |
| Developer tools (CLI, SDK, Terraform) docs.databricks.com | docs.databricks.com | None |
| Feature Store docs.databricks.com | docs.databricks.com | None |
| Platform release notes (AWS) docs.databricks.com | docs.databricks.com | RSS |
| Python SDK releases github.com | github.com | RSS |
| Serverless compute docs.databricks.com | docs.databricks.com | None |
| Serverless environment versions docs.databricks.com | docs.databricks.com | None |
| Terraform provider releases github.com | github.com | RSS |
| Unity AI Gateway docs.databricks.com | docs.databricks.com | None |
| What's coming docs.databricks.com | docs.databricks.com | None |