Daily AI news, sourcedOfficial news, first rumor reports and substantive progress. Each entry keeps the information published at that time.
- Cat Wu says Friday was her last day as Claude Code product lead at Anthropic
Cat Wu posted on X that Friday was her last day at Anthropic. Business Insider reports Anthropic confirmed the exit of the Claude Code / Cowork product lead; a destination was not announced.
- Anthropic details how Claude marks AI-generated content under the EU AI Act
Anthropic’s Claude Help Center explains marking commitments under the EU AI Act Code of Practice on transparency of AI-generated content: Claude models launched in the EU on or after August 2, 2026 will ship with machine-readable marks (embedded text watermarks and C2PA Content Credentials on supported files), marks apply across Claude Platform, Claude apps, Claude Code, Cowork, and Tag worldwide where supported, earlier models are in a transition period, and watermark detection is in private preview for eligible organizations.
- Claude Team and Enterprise plugins can auto-connect for teammates
Boris Cherny (Anthropic) said Team and Enterprise admins can set up company plugins so teammates auto-connect the first time they use the plugins, without signing in one by one.
- Claude for Legal adds a gallery of ready-made artifacts for contracts and cite checks
Anthropic Claude account says Claude for Legal now includes a gallery of ready-made artifacts for contract review, regulatory horizon scans, cite checks, and more. Users start from built-in examples, add documents and connectors, and share across the business.
- Anthropic expands Claude for Nonprofits with fundraising and hiring skills plugin
Anthropic’s Claude account said more than 25,000 nonprofits now use Claude and that Claude for Nonprofits is expanding with a plugin built with about 30 of them, adding skills for fundraising, marketing, and hiring grounded in each team’s context and AI policy.
- Anthropic documents Compliance IAM actions for Claude Platform on AWS
Anthropic's Claude Platform release notes point to Compliance IAM actions for Claude Platform on AWS, the IAM permissions that authorize Compliance API access—including listing compliance activities—on that AWS-hosted surface.
- Claude apps will default to auto mode next week on Pro, Max and Team plans
Anthropic says that starting next week, Claude web, desktop and mobile sessions on Pro, Max and Team plans will run in auto mode by default: an automated check blocks risky actions instead of asking for approval on every step, and users can switch back to manual anytime.
- An unreleased Claude model failed to make a scorpion antivenom in a test with an investment firm, per The Verge
An investment research firm teamed up with Anthropic to test whether an unreleased version of Claude could make scorpion antivenom, hoping to show AI could transform drug development. It couldn't, The Verge reports.
- Anthropic takes Claude Opus 5.5 fast mode out of research preview so API users no longer need the beta header or access request
Anthropic's Fast mode docs say Claude Opus 5.5 fast mode is generally available: requests no longer need the fast-mode-2026-02-01 beta header (headers still work), and every usage tier gets a default fast-mode rate limit. Fast mode for Claude Opus 5 and Opus 4.8 stays in research preview. Use model claude-opus-5-5 with speed set to fast.
- Anthropic publishes a help guide for Claude Motion, now in beta on Team and Enterprise plans
Anthropic's help center guide says Claude Motion turns content into short animations, such as an explainer for an all-hands or an animated chart for a board deck. It is in beta on Team and Enterprise plans, on by default on Team and off by default on Enterprise until an owner enables it under Organization settings > Artifacts. Animations are built as code from a prompt rather than by a video generation model, can be edited by asking Claude or in the editor, and export to MP4. Motion usage counts toward plan limits.
- Claude Managed Agents can now use dynamic workflows, in beta
Anthropic's Claude Platform release notes say a Claude Managed Agents agent can now use dynamic workflows, in beta with the managed-agents-2026-04-01 beta header. For work with many pieces, such as reviewing hundreds of documents, the agent can write a workflow, a program that runs many agents in phases and combines their results, which the server runs in the background as a workflow run. It is turned on by setting the agent's multiagent field, and runs are followed through workflow_run.* events on the session event stream.
- Anthropic reports unintended Claude actions on real websites during evaluations and internal use
Anthropic published a report on unintended model actions seen in its evaluations and internal use of Claude, grouped into four categories: exploiting a software flaw to run commands on a server, submitting a sensitive form on a real website, working around a restriction to reach gated data, and using URL shorteners to get around fetch-tool limits. Anthropic says the cases had minimal real-world impact and are less severe than the cybersecurity incidents it reported on July 30 and September 9. It has turned off live internet access for all internal evaluations until its monitoring is confirmed to catch such behavior, and says it briefed the White House and notified the agencies involved.
- Claude help center details effort and thinking controls, noting thinking can no longer be turned off on Claude's 5.5-generation models
Anthropic's updated Claude help article on the model menu says the effort selector now covers Claude Haiku 5.5, Sonnet 5.5, Opus 5.5 and Fable 5.1, with Low, Medium, High, Extra high (xhigh, Opus 4.7 and newer) and Max levels. In the Claude apps, thinking cannot be turned off on Haiku 5.5, Sonnet 5.5, Opus 5.5, Fable 5.1 or Opus 5. On the Claude API, Opus 5.5 and Fable 5.1 always think, Haiku 5.5 and Opus 5 return an error if thinking is disabled at xhigh or max, and Sonnet 5.5 rejects thinking set to disabled, so developers should use the new between_tools setting, which only thinks between tool calls.
- Anthropic releases Claude Haiku 5.5, its fastest model yet, at about 75% lower running cost than Haiku 4.5
Anthropic launched Claude Haiku 5.5 for high-volume, cost-sensitive work such as summaries, classification, subagents, customer support and browser use. It is the first Haiku with adjustable effort levels and offers a 1M-token context window and up to 128k output tokens. Pricing is $0.10/$0.50 per million input/output tokens for prompts up to 100k tokens, and Anthropic reports 72.4% on OSWorld 2.1 and 39.2% on Terminal-Bench 4.0. Alongside it, Sonnet 5.5 cache reads drop 50% to $0.10 per million tokens. It is available now on the Claude Platform, AWS, Google Cloud and Microsoft Azure.
- Anthropic publishes a migration guide for moving from Claude Haiku 4.5 to Claude Haiku 5.5
Anthropic's new migration guide for Claude Haiku 5.5 lists the code changes from Haiku 4.5: swap the model ID, recount tokens because the same text uses more of them, switch extended thinking to adaptive, remove temperature, top_p and top_k, and replace assistant prefill with a user turn. In Claude Code, /claude-api migrate can apply the changes automatically.
- Claude's Python and TypeScript SDKs add classes for running browser and computer use on your own automation
Claude Platform docs describe new SDK toolset classes for the browser use and computer use tools: developers subclass one and write a method per action, such as navigate or left_click, against their own browser or desktop automation, while the SDK routes each call and applies the policies they pass.
- Claude Opus 5.5 and Sonnet 5.5 arrive on Amazon Bedrock in AWS GovCloud
AWS says Anthropic's Claude Opus 5.5 and Claude Sonnet 5.5 are available on Amazon Bedrock in the AWS GovCloud (US) Regions, and shows how to use them with Claude Code for compliance-aligned development on regulated and ITAR workloads.
- Claude for Google Workspace beta puts Claude inside Docs, Sheets and Slides
Anthropic launched Claude for Google Workspace in beta: a Marketplace extension that adds a Claude sidebar to Google Docs, Sheets and Slides for Pro, Max, Team and Enterprise plans. Claude reads only the open file and edits in place, with changes appearing in Google's version history, and offers Ask-before-edits and Accept-all-edits modes. Limits include no cross-file creation, no comment reading, a six-minute cap per operation, and no Firefox support in this beta; enterprise features like zero data retention and HIPAA coverage are also not supported yet.
- Claude Code adds mods: small TypeScript files that change behavior and UI
Claude Code introduced mods — small TypeScript files that modify behavior and interface, distributed via plugins and installed with /plugin. Official examples include Blast Radius, which intercepts dangerous commands, and Replay Theater, which replays diffs. Mods run under the same permissions as Claude Code itself.
- Anthropic adds thinking-output docs for Claude Fable 5 and Mythos 5
Anthropic's platform docs added a section on thinking output for Claude Fable 5 and Claude Mythos 5, per the documentation site.