AI Daily · Nov 11, 2026 · Wednesday
BAD INFLUENCE
Google releases Gemini 4 Carbon, its strongest model, for engineering work
Google says Gemini 4 Carbon is now available and is its strongest model, with frontier performance in cybersecurity and long-horizon capabilities. The post says Carbon explores, reasons, and runs extended verification loops for complex engineering workloads. Google is also previewing /swarm in Antigravity, which lets one prompt spin up a team of agents with Carbon handling architecture and auditing.
- Google releases Gemini 4 Carbon for general availabilityGemini API Changelog
- Google rolls out Gemini 4 Carbon in Google AntigravityGoogle Antigravity
- Google explains how to build with Gemini 4 CarbonGoogle for Developers
Model releases & updates
Google · Gemini appOfficial12 more sourcesGoogle makes Gemini 4 Carbon available to developers and consumers
Google says Gemini 4 Carbon is now available through the API in Google AI Studio and Gemini for business at $2 per million input tokens and $10 per million output tokens. It is also rolling out to Google AI Ultra subscribers in Google Antigravity, the Gemini app, and AI Mode in Search. Google positions Carbon for tasks that test a model's limits, such as complex coding and scientific research, and says it uses more tokens than Argon to reason and verify.
MiniMax releases open-weight H3-Image model for 4K image generation
MiniMax releases H3-Image, an open-weight image generation model built on H3 that supports text-to-image generation, reference-based generation, and image editing. The company says it can generate a 4K (4096 × 4096) image in under 10 seconds and accept up to 16 reference images. H3-Image is also available through the MiniMax API Platform, MiniMax Code, and Hailuo AI, with 50% off standard pricing for 4K-and-above requests during launch week.
Nagli@galnagliXGoogle makes Gemini 4 Carbon publicly available for security work
Google has made Gemini 4 Carbon publicly available, and the post says it is the state-of-the-art model on ExploitGym and pentesting. Nagli says his team used it to find and help fix critical cross-tenant 0-day vulnerabilities, and the attached image describes Wiz's Scan for Good program, where the model found 0-days in two major SaaS providers in under two hours.
Product launches & updates
Claude BlogOfficial4 more sourcesAnthropic introduces Claude Sheets, Whiteboards, and Data with interactive connectors
Anthropic introduces Claude Sheets, Claude Whiteboards, and Claude Data, which let teams build spreadsheets, whiteboards, and dashboards with Claude. Partners can now embed their apps directly in Claude through interactive connectors, and Claude auto-connect is in beta on eligible Team and Enterprise plans. Claude Meetings, which will take notes and update team documents during calls, is coming soon.
OpenAI · YouTubeOfficial5 more sourcesOpenAI introduces Businesses in ChatGPT in beta with Business Hub and Business Agents
OpenAI introduces Businesses in ChatGPT in beta, letting verified businesses claim their listing in Business Hub and update details and product feeds with current pricing and availability. Eligible U.S. ecommerce and online service businesses can also build a Business Agent that answers customer questions in a conversation labeled separately from ChatGPT's independent answers. OpenAI says it does not share people's conversations or Business Agent chats with businesses.
Google Cloud · AI & Machine LearningOfficial1 more sourceGoogle launches Gemini for healthcare in preview for payers and providers
Google announces Gemini for healthcare, a healthcare-specific version of its Gemini agent, now available in preview in the U.S. It connects to EHR, claims, and coverage systems while keeping existing role-based access controls and HIPAA-oriented governance in place. Google says clinical data is never used to train its foundation models.
Industry news
Anthropic NewsroomOfficial6 more sourcesAnthropic launches roadshow for proposed $70 to $78 per share initial public offering
Anthropic has launched the roadshow for its proposed initial public offering of Class A common stock. The offering consists of 825,000,000 shares, with 801,351,354 sold by Anthropic and 23,648,646 by existing stockholders, at an expected price of $70.00 to $78.00 per share. Anthropic will receive no proceeds from the shares sold by stockholders and has applied to list on Nasdaq under the symbol ANT. The securities may not be sold until the registration statement becomes effective.
Anthropic sets IPO price range of $70 to $78 per share, targeting Nov. 19 Nasdaq debut
Anthropic filed an updated prospectus offering 825 million shares at $70 to $78 each, which could raise up to $64.35 billion at the top of the range. Chipmakers and cloud providers, including Nvidia, Google, and Microsoft, agreed to buy $26 billion of Class N stock in a private placement at the IPO price. The company is reportedly targeting trading on Nasdaq under the ticker ANT on Nov. 19, and the company's stated valuation could reach $1.405 trillion.
Anthropic reports $33.2 billion in nine-month revenue ahead of record IPO
Anthropic reported $33.2 billion in revenue for the first nine months of 2026, up 1,206% from a year earlier, in its updated IPO prospectus. The company plans to offer 825 million shares at $70 to $78 each, which at the high end would raise $64.35 billion at a valuation of up to $1.405 trillion. CNBC's Jim Cramer says the growth rate justifies the deal, while noting that quarter-over-quarter growth slowed to 47% in the third quarter from 144% in the second.
Rohan Paul@rohanpaul_aiX1 more sourceAnthropic says it has crossed a Responsible Scaling Policy threshold in amended S-1
Anthropic says in an amended S-1 filing that it has crossed a Responsible Scaling Policy threshold, where AI is accelerating AI research and stronger safety checks are needed. The company bases the call on what it expects from future models rather than on what current models have shown. The same filing sets the IPO price at $70 to $78 a share, valuing the company at up to $1.4T.
Papers & research
Google DeepMindOfficial6 more sourcesFollow-upGoogle DeepMind uses AI agents to formally verify Signal's cryptographic protocol code
Google DeepMind says researchers used its AlphaProof Code agentic framework to mathematically prove key security properties of the code underlying the Signal protocol. The agents generated over six million lines of Lean and flagged about one percent of the codebase for human review, and the work took four weeks. The team also says it open-sourced its formal verification artifacts.