Releasesby GlyphClaude API adds the anthropic-workspace-id response header
Responses now carry the wrkspc_-prefixed ID of the workspace the API key or access token resolved to, including the Default Workspace, making it easier to attribute traffic to workspaces.
Also shipped Aug 11, 2026Anthropic in August 2026 →
Query templates for data clean rooms are generally available (GA)
Query templates for data clean rooms are generally available (GA). Query templates allow data clean room owners and publishers to share predefined queries without exposing the underlying tables and views. Additionally, table parameters in table-valued functions (TVFs) are generally available (GA).
Databricks SQL version 2026.32 is now available in Preview
Databricks SQL version 2026.32 is now available in the Preview channel. Review the following section to learn about new features and improvements. Validate VARIANT values with is_valid_variant The is_valid_variant function returns true when its input is a well-formed VARIANT value and false otherwise.
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.
REPLACE USING flows for standalone streaming tables (Beta)
REPLACE USING flows are now available for standalone streaming tables in Beta . Use a REPLACE USING flow when your source is a series of partial snapshots keyed by column. On each update, the flow replaces all rows that match the specified key columns and leaves all other rows unchanged.
Unity Catalog ABAC GRANT policies support more securable types in Beta
ABAC GRANT policies dynamically grant Unity Catalog privileges to securables whose governed tags match a condition. These policies were previously only available for models, but now also support model services, model provider services, Model Context Protocol (MCP) services, and agent services. All supported types are in Beta.
Gemini Enterprise: New data stores and support for new actions (Public Preview)
The following data stores are available in Public Preview in Gemini Enterprise: Cisco Workspaces Fibery Gong Guru Hex LegalZoom Mercury Miro Ramp Solve Intelligence Vanta Webex Meetings You can search and read data from these data stores using natural language.
More from Claude PlatformEvery Claude Platform release →
Claude Sonnet 5 introductory pricing of $2 / $10 per MTok becomes permanent
The introductory Claude Sonnet 5 price is now standard, and the scheduled increase to $3 / $15 per MTok on September 1, 2026 will not happen.
Claude Managed Agents sessions support hard spend budgets
A session budget caps spend at public list rates; when reached, the session pauses with the budget_reached stop reason instead of starting new model requests. Deployments can apply the same budget to every session they start.
Claude Managed Agents sessions can consult an advisor model
An advisor entry in the agent's multiagent roster names a model at least as capable as the agent's own that the primary thread can consult mid-turn for strategic guidance.
Claude Managed Agents adds inference_geo to control where inference runs
Set inference_geo in the model object when creating an agent, or override it per session, to control data residency for model inference.
Claude Managed Agents sessions load skills from GitHub repositories
When a session mounts a repository, skills in its root .claude/skills directory are discovered automatically at session start and available to the agent.
Inference hooks
are now in beta for Claude Enterprise organizations. Point Claude at your organization's AI security server, and each governed prompt across claude.ai, Cowork, and Claude Code is held for the server's allow or deny verdict before inference proceeds.
Claude Opus 4.1 is retired on the Claude API
All requests to claude-opus-4-1-20250805 now return an error. Anthropic recommends upgrading to Claude Opus 5; researchers can request continued access through the External Researcher Access Program.
Claude Console Workbench is renamed playground
Playground supports every Messages API parameter, includes templates for features such as code execution and web search, and shows the full SDK request and API response for each run.