Unity Gateway release notes

Unity Gateway is the enterprise control plane for governing AI cost, security, and access across the models, MCP servers, and agents used across your organization.

nota

Releases are staged. Your Databricks account might not be updated with a new feature until a week or more after the initial release date.

Monitor and cap external provider costs

August 28, 2026

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. To turn it on, enable External Model Spend in Budgets (Beta) on the account console Previews page, then select Include external model usage when you create a budget scoped to the Unity Gateway resource type.

See Include external model usage.

Zhipu AI GLM 5.3 Flash is now available on Unity Gateway

August 26, 2026

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.

Customers are responsible for ensuring their compliance with applicable terms.

To query the model and see supported regions, see Use foundation models.

xAI Grok 4.6 is now available on Unity Gateway

August 26, 2026

xAI Grok 4.6 is now available on Unity Gateway for reasoning-heavy coding and agentic 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.

Customers are responsible for ensuring their compliance with applicable terms.

To query the model and see supported regions, see Use foundation models.

Unity Gateway now supports Provisioned Throughput

August 24, 2026

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 AI Gateway.

Monitor all AI activity with the unified trace table (Beta)

August 21, 2026

The unified trace table is in Beta. It captures every request and response across your Unity Gateway services into one Unity Catalog table in OpenTelemetry format, giving you a single place to monitor, debug, secure, and audit all AI activity.

See Monitor all AI activity using the unified trace table.

DeepSeek V4 Pro (0813) is now available on Unity Gateway

August 14, 2026

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.

To query the model and see supported regions, see Use foundation models.

Google Gemini 3.7 Flash is now available on Unity Gateway

August 13, 2026

Google Gemini 3.7 Flash is now available on Unity Gateway for fast, cost-efficient everyday tasks. 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.

Customers are responsible for ensuring their compliance with applicable terms.

To query the model and see supported regions, see Use foundation models.

Block or redact sensitive data in requests and responses

August 13, 2026

The Sensitive Data Detection guardrail (Beta) detects structured sensitive data (such as Social Security numbers, credit card numbers, IBANs, and phone numbers) and either blocks the interaction or redacts the matched values in place. Detection is deterministic (regular expressions plus checksums and context keywords, with no evaluator model), so it adds under 50 ms for a 100-turn conversation and returns the same result every time. To use it, enable the Enhanced Unity Gateway preview on the account console Previews page.

This release also includes the following improvements:

• Blocked interactions now return a standard HTTP 200 response with the block reason, instead of a 4xx error, so coding agents that resend history don't re-trigger the same block every turn.
• Built-in LLM-as-a-judge guardrails can now evaluate a window of recent conversation turns, or the entire conversation history, instead of only the latest message.
• A new execution flow view shows the order in which guardrails run, end to end.

See Detect sensitive data with a service policy.

Smart Routing: Cut AI costs by routing each task to the right model

August 13, 2026

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. Smart Routing works with Claude Code and Codex when you run them through Omnigent or ucode, and routes across both proprietary and open models. Only Omnigent can route across coding harnesses, not just models. In internal benchmarks, it matched frontier-level quality while cutting overall task cost by more than 30%.

See Smart Routing for coding agents.

ABAC GRANT policies now cover Unity Gateway resource types (Beta)

August 11, 2026

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. This lets you grant access to gateway resources by tag instead of one at a time. All supported types are in Beta.

See ABAC GRANT policies.

DeepSeek V4 Flash (0731) is now available on Unity Gateway

August 10, 2026

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.

Customers are responsible for ensuring their compliance with applicable terms.

To query the model and see supported regions, see Use foundation models.

Databricks-managed MCP connectors are now integrated with Unity Gateway (Beta)

August 6, 2026

All Databricks-managed MCP connectors for Genie One and Genie Code have migrated to the Unity Gateway. This brings Databricks-managed connectors under Unity Gateway, enabling centralized governance, access controls, and visibility alongside your other MCP servers and tools.

Users must reauthenticate their connectors to continue using them with Genie One and Genie Code. A notification in Genie prompts impacted users to reconnect.

Moonshot AI Kimi K3 is now available on Unity Gateway

August 6, 2026

Moonshot AI Kimi K3 is now available on Unity Gateway for long-horizon coding, knowledge work, and reasoning, with native vision. 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.

To query the model and see supported regions, see Use foundation models.

Unity Gateway API: Manage resources with SDKs, the CLI, and Terraform

August 6, 2026

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. The API is available across standard Databricks developer tools:

• Python SDK: databricks-sdk 0.125.0 and later
• Go SDK: databricks-sdk-go v0.170.0 and later
• Java SDK: databricks-sdk-java 0.143.0 and later
• Databricks CLI: v1.11.0 and later
• Terraform provider: 1.124.0 and later

See the Unity Gateway API reference.

Unity Gateway is now generally available

August 4, 2026

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.

Some capabilities, including service policies and agent services, remain in Beta.

See AI governance with Unity AI Gateway and AI governance guide.

Use Unity Gateway with your own Claude enterprise subscription

July 30, 2026

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.

Connect and govern any external model provider

July 17, 2026

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.

With external model providers, you get:

• Choice: Adopt any model on day one, and switch between models or OpenAI-compatible providers without re-integrating.
• Control: Govern access with standard GRANT and REVOKE (EXECUTE to query, MANAGE to update), and attach rate limits and guardrails that apply to every request.
• Clarity: Usage and spend tracking is on by default, with token counts, latency, and cost attribution per user, team, or app. Attach inference tables to log full request and response payloads to a governed Delta table.

See Govern external model providers (model provider services).