Made with ♥ in Silicon Valley. Copyright © Joy Larkin 2002-2026.
contact ••• speaking ••• cleverhack.com ••• llms.txt •••
November 1, 2026
The frontier AI labs are getting restless. For years, they've been building models and serving APIs. That's all changing.
OpenAI, Anthropic, Google DeepMind, and Meta have brought to market incredible intelligence capabilities and have captured massive attention. Now they're looking beyond models and APIs and toward actual products people use daily. This makes sense, as everyone is searching for new revenue streams and the lucky have market position to leverage.
I see raw opportunity for products that live in the latent space between AI intelligence and practical end user needs. But here's the kicker — when frontier labs build products, they don't just compete against others, they can quickly collapse entire market segments. This is bad news for a certain class of emergent AI startups and some incumbents who thought their vertical moats were safe.
AI Lab |
Focus |
Likely Products |
|---|---|---|
Anthropic |
Professional productivity |
Advanced coding assistants, writing platforms, intelligence for finance, legal, health tech, and other business verticals |
OpenAI |
Life assistant ecosystem |
Advanced coding tools, AI browser, autonomous agents, voice interfaces, business platform integrations |
Google DeepMind |
Product integration + solving hard problems |
Google product AI embeds, medical/science/biology solutions, image and video understanding/generation, coding tools |
Meta Superintelligence |
Hardware + platform |
Wearables AI integrations, Meta product AI embeds, open source tools |
The wider implication? Many of today's AI startups will need to find new ways to add value quickly when frontier AI labs release competitive products. The next 12-24 months should be interesting to watch.
• Google file editing in the Gemini app, letting you generate, preview, and edit documents, PDFs, slide decks, spreadsheets, and code right inside a chat—typing edits directly or highlighting text for Gemini to revise—available to all users, with PDFs on web and mobile and other formats on web.
• Google textbooks in Gemini Notebook, adding hundreds of free textbooks from OpenStax, Open Textbook Library, and S Chand Publishing, with answers grounded in the textbook and cited inline.
• Google Google AI for Healthcare Professionals Certificate, a self-paced, five-hour course built by Google clinicians on using AI for everyday administrative work, offering up to 5.0 AMA PRA Category 1 Credits.
• Google deepens its health partnership with Singapore, opening a Google Research hub there, bringing open models like MedGemma into frontline care, and working with NHG Health on an AI health coach.
• Anthropic Usage and Cost Analytics API in beta, reporting token usage and cost for Claude Console and Claude Enterprise organizations by model, workspace, API key, or user, and replacing the older Usage and Cost API.
• Google Gemini 4 Carbon, its strongest model, built for cybersecurity defense and long-horizon agentic work, arriving two weeks after Gemini 4 Argon and ranking #1 on the Artificial Analysis Intelligence Index v5.0. It is going first to a subset of cyber defenders, with access widening to trusted partners through Google's Fairwind Program, at an introductory price of $2/$10 per million input/output tokens.
• Anthropic expands AI for Science to foundations and research institutions, naming 20 scientific funders, institutions, and research groups worldwide as partners in its program of free Claude credits for high-impact research.
• Google Connected Apps Platform in preview, giving developers one place to build, register, and submit apps for both Search and the Gemini app via MCP, alongside new Connected Apps from Adobe, Calm, and Peloton.
• Google immersive explorations in the Gemini app, a library of interactive, layer-by-layer visual explainers, from jet engines to the Silk Road, in the Gemini student hub.
• Google expands ads in AI Mode to Australia, Canada, India, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, the Philippines, and Singapore, and launches multiturn ads in follow-up AI Mode queries in the US.
• Google enhanced autofill from Gmail in Chrome, using Gemini with Personal Intelligence to suggest order numbers, tracking details, and flight bookings from your inbox as you fill out forms, rolling out in the US.
• Meta Muse launches in Singapore, with connectors to Gmail, Google Calendar, and Instagram and local partners including Grab, Shopee, Singapore Airlines, and Lazada.
• Meta partners with five US universities (Caltech, Ohio State, LSU, Indiana University, and UTHealth Houston), granting $2 million and model access for research in medicine and science.
• Google Genie 4, its new world model for real-time interactive simulation with native audio, letting you steer scenes in natural language. It is rolling out to a limited group of trusted testers first, and is coming soon to Google AI Ultra subscribers in Project Genie and to early-access partners in the Gemini API.
• OpenAI expands its partnership with Oracle at Oracle AI World. The two companies announced a product suite that includes Oracle AI Workspace, which links Oracle apps and data to ChatGPT, and Oracle AI Security Mesh, built on OpenAI's cyber models. US customers can buy ChatGPT Work, Codex, and the OpenAI API through Oracle Marketplace now.
• OpenAI iCloud Mail and Calendar plugins for ChatGPT, letting ChatGPT, and Dots, search and send iCloud email and read, schedule, and update iCloud Calendar events, on all plans.
• Anthropic expands its GSA partnership with a 40% discount on Claude Enterprise, Claude for Government, and the Claude API across every purchasing route, extended to state, local, tribal, and territorial governments, plus a new program to help agencies put Claude to work. Anthropic estimates the savings at $1 billion.
• OpenAI GPT-6.1 Terra, a fast, efficient model for visual reasoning and computer use, available in ChatGPT Work and Codex and in the API at $1/$5 per million input/output tokens, half the price of GPT-5.6 Terra and GPT-6.1 Sol.
• OpenAI text your dot, letting ChatGPT Pro users in the US reach their dot over iMessage or RCS with text, photos, and voice messages, in beta.
• Anthropic Claude Tag API in beta for Claude Enterprise, letting admins read and change Claude Tag settings on Slack channels and workspaces programmatically.
• Anthropic agrees to back a European AI consultancy with Hg, the London private equity firm, according to the Financial Times—an independent company expected to launch by the end of 2026 with hubs in London and Berlin, embedding consultants and engineers inside clients to build on Claude. It's Anthropic's first such move in Europe, following its US enterprise services firm in May.
• Meta Admin AI on Forum, which flags pending posts that break group rules and helps admins draft posts in Forum, its app for Facebook Groups, alongside a more flexible Admin Assist for routine moderation.
• Anthropic Claude Models on Google Cloud (Anthropic managed), now generally available, gives Google Cloud customers the full Claude Platform—with new features shipping the same day as on the native API—through Google Cloud IAM, a single Google Cloud invoice, and existing Google Cloud commitments, landing a day after OpenAI's Google Cloud news.
• Anthropic and OpenAI donate Agent Skills and Open Responses to the Agentic AI Foundation, placing Anthropic's skills standard—already integrated by more than 70 agent platforms including ChatGPT, Cursor, GitHub Copilot, and VS Code—and OpenAI's Open Responses under the same Linux Foundation umbrella as MCP.
• OpenAI makes the Decisions API widely available about two weeks after its public beta, returning choices and confidence scores in 100 to 180 milliseconds as it competes with TypeSafe's Jev decision model.
• Anthropic Claude Duet, a research preview of a proactive agent app built around Claude Opus 5.5 that sends a morning brief across your connected apps, messages you only when something needs attention, keeps working when you step away, and checks with you before acting. It's on iOS, Android, and claude.ai for Pro and Max subscribers via a waitlist.
• OpenAI models are coming to Google Cloud, with GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna rolling out in private preview to select enterprises in Model Garden and payable through Google Cloud spending commitments, plus Codex workflows and, soon, OpenAI models in Antigravity—six months after OpenAI came to AWS.
• OpenAI acquires Orchestra Bio, a startup whose AI agents monitor studies, track drug development timelines, and flag delays, folding its team and technology into Rosalind, OpenAI's life sciences suite alongside GPT-Rosalind. Terms weren't disclosed.
• Google new Gemini tools for small business owners, a dedicated business hub that connects to Shopify, Canva, and Google Workspace, generates on-brand visuals from uploaded brand guidelines, and turns messy spreadsheets into financial insights.
• OpenAI locked chats in ChatGPT, protected by a PIN, Face ID, or Touch ID and kept out of regular history and search, though locking doesn't add end-to-end encryption.
• Google Personal Intelligence for Gemini in cars, letting Gemini on Android Auto and cars with Google built-in pull from your email, photos, and calendar for hands-free, tailored answers.
• Anthropic user profiles in the Claude API, letting platforms, gateways, and resellers tell Anthropic which end customer each request is for so safety enforcement can target that customer, as inference hooks leave beta for Claude Enterprise.
• Google co-creation in Gemini Notebook, letting you review and steer the outline for Slide Decks and the producer notes for Video Overviews before they're generated.
• OpenAI ChatGPT Slides, decks built from a conversation, files, or an outline inside Space, editable directly or through ChatGPT and exportable to PowerPoint, PDF, or Google Slides—rolling out to Pro, Business, and Enterprise.
• OpenAI deeper Chrome integration for ChatGPT, letting the extension answer from your open tabs, act on them to send messages, fill out forms, and download files, and help organize them from a new sidebar.
• OpenAI Rosalind Academic Plan, giving approved US university labs up to 30 free ChatGPT Business seats for up to 12 months, with GPT-Rosalind access subject to trusted-access approval.
• Google more ways to connect with the web from Search, adding more inline links, a section for official websites, news carousels, and creator profiles to AI Mode and AI Overviews.
• Google scam checks for text messages with Circle to Search on Android.
• Anthropic OpenTelemetry export, a beta for tracing Messages API requests and Claude Managed Agents sessions in your own observability platform.
• Anthropic licenses Opaque's confidential AI technology and hires CEO Aaron Fulkerson, CTO Rishabh Poddar, and some engineers as Opaque winds down.
• Google Gemini 4 Argon for robotics, scoring 87% on the Embodied Reasoning benchmark, ahead of every other frontier model, with developer prompts for robot control, simulation twins, and physical understanding—including steering a Boston Dynamics Spot.
• Meta Your algorithm on Facebook, AI-powered tools that let people describe in their own words what they want in Feed and Reels, with recommendations adapting in real time.
• Google Gemini 4 Argon is now available for everyone, two weeks after a launch limited to cyber defenders—globally through the API for developers and enterprises, in Antigravity for Google AI Ultra subscribers, and in the Gemini app and AI Mode in Search for Ultra and Pro subscribers. It debuts at #1 on the Vals Index, which covers finance, coding, legal, and tax work, at an introductory $2 in and $10 out per million tokens.
• Meta Muse Wonder and Muse Wonder Full, Meta's most capable models, built for personal and professional agentic tasks and pairing Active Memory with Active Focus to track what matters to users and work toward their goals. Muse Wonder is available now through the Meta Model API and Muse Code, Wonder Full is open to Muse Code subscribers, and both come to the Muse app soon, starting with subscribers.
• Anthropic HCLTech global premier partnership, to scale governed agentic AI across HCLTech's global enterprise clients.
• Anthropic Cyber Mission, launching the Critical Infrastructure Defense Program, which brings frontier Claude models, on-site engineers, and threat research to the providers that defend power, water, and transportation systems, with founding partners including CrowdStrike, Palo Alto Networks, Dragos, and Accenture. It also launches OSS Scanner, which gives open-source projects regular security scans from Anthropic's strongest models for free.
• Google Gemini agent, pitched as a universal agent for work that plans tasks, uses skills and tools, writes and runs code, and returns finished output inside Gmail, Drive, Docs, Sheets, and Calendar. It connects to company systems and any MCP server, and industry versions for financial services and legal are in preview now.
• Anthropic Claude Dashboards and Claude Motion, live dashboards built from connected data sources (beta on paid plans) and code-based animations with MP4 export (beta on Team and Enterprise), as Claude Docs, Slides, and Design reach general availability on every plan, including Free.
• OpenAI GPT-6 comes to ChatGPT with Intelligent UI, so responses can include interactive visuals and in-chat tools like calculators and bill splitters, and start streaming while the model is still reasoning. GPT-6 Sol goes to Plus, Pro, Business, and Enterprise, and GPT-6 Luna to Free and Go users.
• Anthropic Claude Haiku 5.5, a small model for fast, high-volume work with a 1M-token context window, priced at $0.10 in and $0.50 out for prompts up to 100K tokens. That matches GPT-6 Luna's base rate and is about 75% cheaper on average than Haiku 4.5 once tokenizer changes are counted.
• Google improved SynthID Detector, its tool for checking whether content was made with Google AI, now available globally.
• OpenAI teen learning tools and College Planner, adding flashcards and quizzes for teens plus a coming-soon college application planner for US students.
• Anthropic monthly API credits for Max and Team plans, for running your own apps and agents on the Claude Platform.
• Anthropic SDK toolsets for browser use and computer use, beta support in the Python and TypeScript SDKs.
• Anthropic expands its Cyber Verification Program, merging Project Glasswing into one program with three tiers: Defense Access for incident response and vulnerability validation, Red Team Access for authorized penetration testing, and Specialized Access for testing critical systems like power grids and interbank infrastructure. Every tier covers Mythos 5.1, Opus 5.5, and Sonnet 5.5.
• Anthropic Claude for Google Workspace, a beta add-on that puts a Claude sidebar inside Google Docs, Sheets, and Slides to edit the open file in place: rewriting text, building formulas, pivot tables, and charts, and generating slides from existing layouts. It's available on all paid plans, alongside new Google file connectors in Claude chat.
• Google Nano Banana 2.1, an update to Nano Banana 2 with better prompt adherence, text rendering, and character consistency across edits at Flash-level speed and cost, generally available in the Gemini API.
• OpenAI Decisions API, a public beta that returns probabilities, choices, or rubric scores from text and images using GPT-6 Luna.
• OpenAI audio uploads in ChatGPT, turning uploaded recordings into transcripts and summaries, with questions about the audio on paid plans.
• Google EmbeddingGemma 2, an open, lightweight multimodal embedding model that maps text, images, audio, and video into one space for on-device use.
• Google Antigravity Boost and Teamwork in Gemini Enterprise, with Boost's multi-agent hierarchy for hard debugging now generally available and Teamwork's autonomous subagents in preview.
• Anthropic expands the Claude Startups program, offering founders up to $7,000 in products and credits.
• OpenAI textGrain, an invisible statistical text watermark, opt-in for API models and rolling out to ChatGPT and Codex in the EU.
• OpenAI new ChatGPT ad format and measurement tools, including a test of labeled visual ads alongside image generation in the US.
• Google skills in Gemini, reusable prompts that will eventually replace Gems, rolling out in Workspace now and in the Gemini app from October 13.
• OpenAI Finances in ChatGPT comes to Free and Go users in the US, letting anyone connect bank, credit-card, brokerage, and credit accounts through Plaid and Experian to get answers about spending, savings, and investments grounded in their own data. It started as a Pro preview in May.
• Anthropic mods for Claude Code, small TypeScript functions that hook into Claude Code to rewrite prompts, block risky commands, add custom UI, or replace built-in features. Mods ship as plugins in the CLI and desktop app, and Anthropic rebuilt its own diff pane and AGENTS.md support as mods at launch.
• OpenAI virtual try-on in ChatGPT shopping, for apparel and accessories, plus product Favorites and folders.
• Google Gemini 4 Argon, the first Gemini 4 model, built for long-horizon work in software engineering, finance, law, and cybersecurity with a 1-million-token output limit. At launch it is limited to vetted cyber defenders in the Fairwind Program, with paid API customers and Google AI Ultra subscribers promised access next at an introductory $2 in and $10 out.
• Anthropic Claude for Government reaches general availability for US federal and state agencies in a FedRAMP High authorized environment, with usage-based pricing, no seat fees, and department-level spending caps. Claude Code CLI and Claude for Microsoft 365 enter early access in the same environment.
• OpenAI Dots and ChatGPT Space, always-on GPT-6 Astra agents that each run their own cloud computer and browser, connect to 4,000+ apps, and work 24/7—reachable through ChatGPT, Slack, and Teams, paired with Space, a shared workspace where people, ChatGPT, Codex, and Dots collaborate on a new agent-native document type called Pages. One Dot per user at launch, on Pro and Business Premium.
• OpenAI GPT-6.1 Sol, approaching GPT-6 Astra on coding, computer use, and professional work at one-fifth its token prices—$2 in and $10 out versus Astra's $10 and $50—shipping in ChatGPT Work, Codex, and the API alongside a new Ultrafast premium speed tier running about 300 tokens per second.
• OpenAI Codex Cloud, making Codex environments persistent and reusable across desktop, browser, and phone so a task started at a laptop can be checked and approved from a train—plus voice instructions in the CLI and an /agents view for tracking parallel work.
• Meta Muse for Small Business, connecting the Muse agent to Asana, Zoom, Intuit, Box, Canva, and Slack plus Meta ad accounts and professional Instagram and Facebook profiles—give it a goal like finding new customers and it learns what the business sells and how the brand sounds.
• OpenAI Sign in with ChatGPT, plugin extensions, and Pro 500, DevDay launches that let people use their ChatGPT plan in other apps and sites, give developer plugins their own sidebar destinations and interactive panels, and add a $500-a-month tier with the highest usage allowance and Ultrafast access. Developers also get computer use in the Agents API and Codex Security in the cloud.
• Anthropic Claude Sonnet 5.5, the fastest Sonnet yet at over 30% higher output speed, using far fewer tokens for the same work to land roughly 30% cheaper per task—at unchanged Sonnet 5 pricing of $2 in and $10 out.
• Meta Meta Enterprise Platform, bringing Meta's AI stack—Muse, Meta Business Agent, the Muse API, and Muse Code—to businesses and developers.
• Anthropic plugin developer portal, where developers submit plugins to the Claude directory, track review, and see usage analytics.
• Anthropic personal connectors in Claude Tag, letting Claude use your own connected tools for Slack channel requests, with review before posting.
• Meta Connect 2026: the Meta Model API goes generally available, camera-free Ray-Ban Meta Audio glasses are set to ship October 13, the Muse Charm device is due in December, and Muse is coming to AI glasses in the coming months.
• Anthropic Claude Marketplace, a place for teams to discover plugins, connectors, agents, and service partners for Claude.
• Meta Muse Realtime Avatar, pairing Muse Realtime Voice with synchronized video for live conversations, watermarked with Meta Video Seal.
• OpenAI ChatGPT Voice upgrade, adding plugins, GPT-6 models, and voice inside ChatGPT Work.
• Google Gemini 3.8 Flash TTS and 3.8 Flash-Lite TTS, its most expressive text-to-speech models, which generate custom voices and direct scene dialogue from text prompts.
• Google new Connected Apps in Gemini, including Airtable, Linear, monday.com, Adobe, Squarespace, Peloton, and Experian.
• Anthropic Claude Opus 5.5, surpassing Fable 5.1 on agentic coding, knowledge work, and computer use while costing 40% less to run at $4 in and $20 out, with 30% faster output and a 20% bump to usage limits on Pro, Max, Team, and Enterprise.
• OpenAI GPT-6 Sol and GPT-6 Luna, built for professional work, coding, and computer use at half their predecessors' rates—Sol at $2 in and $10 out, Luna at $0.10 and $0.50—landing the same day as Opus 5.5 in the first price war of the GPT-6 era.
• Meta Muse for Mac, the first version of Muse that acts directly on your computer—working inside native Files, Mail, Calendar, Notes, and Messages to organize downloads, find lost files, and summarize threads, with opt-in permissions and approval required for sensitive actions.
• OpenAI Astra for Law, GPT-6 Astra paired with OpenAI's own legal search index spanning 230M+ URLs of US case law, statutes, and regulations—covering 99.9% of published precedential case law, offered to select firms via Trusted Access with 26 partner plugins including Thomson Reuters and Harvey.
• OpenAI ChatGPT for Word, a sidebar add-in for drafting, revising, and proofreading inside Word documents, available on every plan including Free and joining the Excel and PowerPoint add-ins.
• Anthropic Life Sciences Verification Program, giving approved research teams beta access to Claude with more permissive biology safeguards—a life-sciences counterpart to its Cyber Verification Program.
• OpenAI Appshots, which shares what's on your screen with the ChatGPT desktop app so you don't have to describe it.
• Anthropic Claude Code Projects, a beta for selected Pro and Max subscribers using cloud sessions.
• Google CC for groups, turning the Google Labs personal agent into a group agent with its own email address for coordinating family calendars and to-dos.
• Anthropic Claude Docs and Claude Slides, bringing editable documents and presentation generation directly into the conversation—private by default, shareable with colleagues, exportable to Google Docs or Word—as Anthropic folds Claude chat and Cowork into a single interface.
• Google opens Google Home to any AI agent with an early-access MCP server, letting Claude, ChatGPT, or another agent of the user's choosing control connected devices, review camera summaries, and query smart home history—requires a Google Home Premium Advanced subscription.
• Meta Muse Code for Windows, running natively without WSL.
• Google Gemini 3.8 Live and 3.8 Live Extended Thinking, live dialogue models built for natural conversation across the Gemini API, Search Live, and Workspace—Extended Thinking reasons and synthesizes speech simultaneously and switches automatically across 97 languages.
• Anthropic Claude for Small Business expands to 43 workflows and 27 new app integrations including Shopify, Salesforce, Stripe, and Gusto, with every automated workflow defaulting to human approval before execution. It's aimed at a segment that drives 44% of US GDP but lags enterprises in AI adoption.
• Anthropic Salesforce in Claude, a beta plugin that brings accounts, opportunities, and pipeline into Claude with 37 pre-built sales skills.
• Google Gemini Notebook study tools, adding flashcards, quizzes, and 60-second Short Video Overviews in over 80 languages.
• Anthropic Claude for Financial Advisors, a free Cowork plugin with skills for pre-meeting prep, portfolio rebalance review, compliance review, and estate and tax briefings, plus connectors to Schwab, BlackRock, Addepar, Envestnet, Orion, and Vanguard.
• OpenAI ChatGPT for Financial Services, a ChatGPT Work configuration built with Morgan Stanley and Evercore that pairs GPT-6 Astra with Daloopa, PitchBook, and LSEG data to build research, models, and client materials—aimed squarely at junior banker workflows.
• OpenAI GPT-Live-1 and the Agents API, bringing full-duplex voice—listening and speaking at once, handling interruptions and backchannels—to developers at $0.05 per minute, alongside public beta access to the same agent harness that runs Codex, with managed sandboxes.
• Google Gemini app for Windows, a native desktop app that brings Gemini to the PC.
• OpenAI Data agent for ChatGPT Work, connecting Snowflake, BigQuery, Databricks, and other sources to analysis and dashboards under existing permissions.
• Anthropic smart reports, a beta that analyzes how a team uses Claude, what the work costs, and where sessions run into friction.
• Meta Muse, a consumer AI agent that books travel, sends emails, shops, and fills out forms on users' behalf—running tasks in a dedicated Muse Secure VM, available in the US via muse.ai, mobile apps, and WhatsApp, free with $20 and $100/month paid tiers.
• OpenAI ChatGPT Images 2.5, adding a Sketch tool that turns rough drawings into finished images and cutting generation latency by up to 50%, alongside two new image models in the API.
• OpenAI ChatGPT Work writing style, which learns how you write from connected tools and carries it into new drafts.
• Google Lyria 3.5 in the Gemini app, its most expressive music model, also in AI Studio, Flow Music, and Vids.
• OpenAI GPT-6 Astra, billed as "the most intelligent and aligned model in the world" and the first to trigger OpenAI's critical-cyber safeguard threshold—full capabilities go to trusted partners, while paid users get a restricted version that declines certain cybersecurity prompts.
• Google WeatherNext 3, its most accurate weather model, now built into Search, Gemini, Maps, and Cloud.
• Google Gemini 3.8 Flash and 3.8 Flash Cyber, its third Flash release in six weeks—approaching frontier-model performance at $0.75 per million input tokens, with the Cyber variant limited to trusted testers via the new Fairwind Program.
• Meta Muse Spark 1.3, built for longer agentic workflows in Muse Code and the Meta Model API with stronger prompt-injection resistance, using about 20% fewer tool calls and 25% fewer tokens than 1.2.
• Anthropic Claude Fable 5.1 and Claude Mythos 5.1, its most advanced models for coding and knowledge work, cutting cache-read fees by 75%; Mythos 5.1 stays behind trusted-access programs for cybersecurity and life-sciences work.
• Meta Muse Voice Transcribe, a real-time speech model with streaming recognition and speaker separation across 70+ languages, in the Meta Model API, Muse Code, and Meta AI for Mac.
• Anthropic Enterprise Frontier Safeguards, a phased system for customer-controlled activity storage and automated misuse monitoring.
• OpenAI ChatGPT for Healthcare adds Epic, plus a Healthcare Public Data plugin drawing on nine official sources.
• OpenAI expands ChatGPT Ads globally, opening direct purchase via Ads Manager across India, Europe, the Middle East, and North Africa after the ad product hit a $1 billion annualized run rate in under 200 days.
• Anthropic Claude for Teachers comes to schools and districts as a free Enterprise offering—administrators verify once, connect SSO, and every educator on the domain gets the teaching skills, state standards, and K-12 privacy terms under central management.
• Google Gemini Omni 1.1 Flash, a generally available video generation and editing model that adds clip extension, first-and-last-frame interpolation, and output up to 4K.
• Google Expert Intelligence, a cross-Google initiative that brings insights from leading authors, publications, and domain experts into Gemini Notebook, starting with ebooks you've bought on Google Play Books.
• Anthropic Model Hardware Standard, a research-preview spec for AI agents that operate lab and manufacturing equipment.
• Anthropic and Salesforce announce Claudeforce, embedding Claude's reasoning in Salesforce's data, workflows, and governance—in pilot now, with open beta expected in September.
• Meta Muse Image in the Meta Model API, priced at $0.01 per image.
• Google Gemini 3.5 Transcribe, a real-time speech-to-text model for developers.
• Google Gemini Live productivity features, moving Gemini Live beyond conversation to handle tasks on your behalf.
• OpenAI ChatGPT Business Premium seats, with five times Standard usage and no five-hour limit.
• OpenAI scheduled tasks upgrade, letting tasks respond to Slack, Gmail, and GitHub changes and bringing tasks to Free users.
• Anthropic memory across chat and Cowork, with every remembered item listed as a topic you can edit or delete.
• Anthropic brings Claude Mythos 5 to more defenders: enterprise customers can now scan their own codebases for vulnerabilities in Claude Security, partner security products get Mythos-powered findings, and a $35M Defender Advantage Fund backs open-source defense.
• OpenAI cuts GPT-5.6 Sol prices by more than 20% for three months.
• Anthropic makes computer use, the browser use tool, the Skills API, and the Files API generally available on the Claude Platform.
• Anthropic Claude Academy, with courses, tutorials, learning paths, and badges.
• OpenAI ChatGPT for Teens, with added safety measures and a Study Mode built to make teens work through problems rather than get answers.
• Anthropic Gmail and Google Drive write actions, letting Claude act in both apps with approval required by default.
• Anthropic Claude Tag for CI/CD incidents, making Claude the first responder to pipeline failures.
• Google Gemini in Chrome auto browse on Android, for AI Pro and Ultra subscribers in the US.
• Anthropic auto mode becomes the default in Claude Code, for Pro, Max, and Team, so sessions proceed unless an action looks irreversible or destructive.
• Google Gemini 3.7 Flash, its most intelligent workhorse model for coding and agents, arriving three weeks after 3.6 Flash at an introductory half price and powering Gemini Spark—while Gemini 3.5 Pro stays delayed.
• OpenAI Computer History, an optional macOS feature that lets ChatGPT and Codex reference selected app and web activity.
• OpenAI GPT-5.6 Sol Ultrafast preview, running Sol at up to 14 times the speed.
• Anthropic Cowork in the Claude in Chrome side panel, sharing sessions, conversations, skills, and connectors across devices.
• Google Pixel 11 series, designed around Gemini Intelligence.
• OpenAI ChatGPT for Linux, bringing ChatGPT, ChatGPT Work, and Codex to Linux in preview.
• Anthropic Compliance API for Cowork and Claude Code, extending the beta beyond chat.
• OpenAI expands Daybreak into Blue and Red tiers, adding GPT-5.6-Cyber—a purpose-trained cybersecurity model for vulnerability research and exploit validation, available only to vetted Daybreak Red participants.
• Meta Muse Glimmer, Meta Superintelligence Labs' first open-weights model—a 30B agentic model under Apache 2.0, distilled from Muse Spark and built to run always-on local agents on a single consumer GPU or Mac.
• OpenAI updated GPT-5.6 Sol in ChatGPT, with GPT-5.6 Luna becoming the default for Free and Go users.
• Anthropic self-hosted Claude Code environments, a Team and Enterprise beta that runs sessions on an organization's own infrastructure.
• Anthropic skill and plugin security scanning, an Enterprise beta that checks third-party skills and plugins for malicious content.
• Meta Muse Code, a terminal coding agent for macOS and Linux powered by the new Muse Spark 1.2, coordinating persistent background agents across large codebases—Meta's answer to Claude Code and Codex.
• Anthropic inference hooks for Claude Enterprise, letting a security server inspect and block prompts and tool responses.
• OpenAI cuts GPT-5.6 Luna and Terra prices, by 80% and 20%, and adds a faster Sol option for API customers.
• Google Gemini Robotics ER 2, its most capable embodied reasoning model, with real-time video understanding and multi-robot collaboration.
• Google Gemini Spark web errands, letting Spark use your signed-in accounts and saved passwords, with permission, to handle tasks like booking viewings.
• Google Lyria 3.5, its newest music model, powering Google Flow Music.
• OpenAI GPT-Live-Transcribe and GPT-Transcribe, transcription models for live audio and batch files.
• OpenAI open-sources the Codex Security CLI and SDK, tools for finding, validating, and fixing vulnerabilities.
• Anthropic brings the MCP 2026-07-28 spec to Claude, with a stateless core, MCP Apps and Tasks extensions, and stronger OAuth.
• Anthropic Claude Opus 5, a new general-purpose flagship with a 1M-token context window and a low/medium/high effort toggle that lets users trade cost against capability—at roughly half the price of comparable frontier models.
• OpenAI voice computer control, letting you control your computer and direct agents in ChatGPT Work or Codex by voice.
• OpenAI OpenAI Presence, a managed enterprise platform for governed voice and chat agents with testing, permissions, monitoring, and human escalation, deployed with OpenAI's forward-deployed engineers and partners. OpenAI says it now resolves 75% of its own inbound support without a human.
• Google Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, three efficiency-focused models for production agents—3.6 Flash uses up to 17% fewer tokens than 3.5 Flash, and 3.5 Flash Cyber is limited to governments and trusted partners—while Gemini 3.5 Pro, promised for June, stays in testing.
• Google Gemini Notebook, NotebookLM extended into the Gemini app and Google Search, with a secure cloud computer for running code.
• Google connected apps in AI Mode, letting you link apps like YouTube Music directly to Search.
• Google Gemini Omni and personal avatars in Google Vids, for generating and editing clips in plain language.
• Anthropic Claude for Teachers, giving verified US K-12 educators a free year of premium Claude—including Claude Code and Cowork—plus teaching skills built with learning scientists and a Learning Commons connector mapped to academic standards in all 50 states, answering OpenAI's ChatGPT for Teachers in the race for the classroom.
• Anthropic self-serve HIPAA configuration, for Claude Enterprise and the Claude Platform.
• Anthropic rebuilt memory, stored as individual, categorized entries Claude reads and updates during conversations.
• OpenAI ChatGPT Work, a GPT-5.6-powered agent that turns a goal into finished sheets, slides, docs, and sites—paired with a unified desktop app for Mac and Windows that merges Codex directly into ChatGPT, letting users switch between chat, Work, and Codex in one app. GPT-5.6 also reaches general availability.
• Meta Muse Spark 1.1 and the Meta Model API, Meta's first paid model API, opening a multimodal agentic model with a 1M-token context window to US developers in public preview at $1.25 in and $4.25 out per million tokens.
• Anthropic Reflect, a monthly recap of the topics you spent time on with Claude.
• OpenAI plugin directory, replacing the ChatGPT apps directory as the way to find capabilities across ChatGPT and Codex, as group chats are retired.
• Google AlphaEvolve, its Gemini-powered code-optimization agent, becomes generally available to Google Cloud customers.
• OpenAI GPT-Live, a new generation of voice models built for more natural conversation, now powering ChatGPT Voice.
• Meta Muse Image, the second major release from Meta Superintelligence Labs after April's Muse Spark, an image generation model aimed at creators and advertisers, with Muse Video planned next.
• Anthropic Claude Cowork comes to mobile, letting users direct long-running agentic work from their phones.
• Meta Muse Video, an early preview of Meta's video model announced alongside Muse Image, coming soon to creators and Meta AI.
• Anthropic write tools for the Microsoft 365 connector, so Claude can draft and send email and manage calendar events.
• OpenAI ChatGPT for PowerPoint, generally available worldwide on all plans.
• OpenAI GPT-Realtime-2.1-mini, adding reasoning and tool use to the Realtime mini lineup at the same price.
• Anthropic redeploys Claude Fable 5 globally after the Commerce Department withdraws its June export-control order—ending a 19-day shutdown that taught enterprises hard lessons about frontier AI as infrastructure. Mythos 5 remains limited to approved users.
• Anthropic Claude Science, an AI workbench for scientists—databases, coding tools, compute, and research workflows in one workspace, pre-configured with 60+ scientific connectors for genomics, proteomics, and cheminformatics. Anthropic will also use it to pursue its own drug programs for rare and neglected diseases. Available in beta to all paid Claude plans.
• Anthropic Claude Sonnet 5, its most agentic Sonnet yet, with substantial gains over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work.
• Google Nano Banana 2 Lite and Gemini Omni Flash, its fastest, most cost-efficient Gemini image model plus a public API preview of Omni Flash, a natively multimodal model for building video workflows.
• OpenAI previews GPT-5.6 Sol, Terra, and Luna, its strongest models yet—initially restricted to "trusted partners" at the US government's request, with general availability promised in the coming weeks.
• Google study notebooks in the Gemini app, for turning class notes and goals into a study plan.
• OpenAI and Broadcom reveal Jalapeño, their first co-designed LLM-optimized inference chip.
• Google computer use in Gemini 3.5 Flash, bringing agentic screen control to its fastest model tier.
• OpenAI GPT-5.5 Instant update, better at inferring intent and adapting responses.
• Anthropic Claude Tag, replacing its Slack app with a persistent AI teammate that learns, monitors, and works autonomously inside team channels.
• OpenAI new Scheduled tasks page, rolling out to Go and paid plans.
• Google Google Home Speaker, a smart speaker built for Gemini.
• Anthropic edit Claude's drafts in place, opening reports and plans beside the chat for targeted edits.
• Anthropic Claude Fable 5 and Claude Mythos 5, the first Mythos-class models: Fable 5 is the most capable generally available model, while Mythos 5 lifts select safeguards for approved cyberdefenders via Project Glasswing with the US government. Days later, a US export-control order forces both models offline.
• Google Gemini 3.5 Live Translate, a speech-to-speech translation model that detects 70+ languages and keeps the speaker's intonation.
• Google NotebookLM upgrade, adding advanced reasoning, a secure cloud computer, and charts, spreadsheets, and slide decks.
• OpenAI Dreaming, a scalable system for synthesizing ChatGPT memory.
• Google Gemma 4 12B, an open model that runs smart AI agents locally on a laptop with 16GB of memory.
• OpenAI GPT-Rosalind update, pairing GPT-5.5's agentic skills with stronger drug-discovery performance as the model leaves research preview for eligible organizations.
• OpenAI Codex for every role, tool, and workflow, with role-specific plugins, interactive Sites for sharing work via URL, and annotations—pushing Codex beyond engineering.
• Anthropic Claude Opus 4.8, most capable generally available model with a 1M token context window by default, stronger agentic judgment, and the highest score on their Legal Agent Benchmark. Also announces Series H raise at $965B valuation with $47B run-rate revenue.
• Anthropic's enterprise services firm acquires Fractional AI, an applied AI implementation company, as its operational core—ending Fractional's 11-month partnership with OpenAI. Anthropic moves into direct competition with the traditional consulting industry.
• Anthropic KPMG global alliance, Claude embedded into KPMG's Digital Gateway platform and available to all 276,000+ employees across 138 countries. Anthropic names KPMG a preferred partner for private equity, with joint products planned for PE portfolio companies.
• Google I/O 2026: Gemini 3.5 Flash, Gemini Spark (agentic assistant across Search, Android, and Google apps), and Omni, a world model for simulating physical environments.
• Google Search at I/O 2026, adding agentic coding in Search alongside Universal Cart, Android Halo, and new Android XR eyewear.
• Google Project Genie expands, letting people simulate real-world places.
• Google AI content identification tools, to help people tell whether content online is AI-generated.
• Anthropic acquires Stainless, the SDK and MCP server tooling company behind every official Anthropic SDK since day one—bringing SDK generation and MCP connectivity in-house as Claude Platform scales.
• Anthropic expands PwC partnership, rolling out Claude Code and Cowork to PwC's global workforce with a joint Center of Excellence and certification program for 30,000 professionals.
• OpenAI personal finance in ChatGPT, connect your financial accounts and ask ChatGPT questions grounded in your actual financial data.
• OpenAI Codex in the ChatGPT mobile app, a preview on iOS and Android that runs tasks in the cloud or on a paired computer.
• Anthropic Claude for Small Business, 15 pre-built agentic workflows inside Cowork with connectors to QuickBooks, HubSpot, PayPal, Canva, and Google Workspace—Anthropic's first product aimed at SMBs rather than enterprise.
• Anthropic Claude Platform on AWS launches, giving developers native Anthropic API access through AWS with day-one feature parity—separate from Bedrock.
• Google Gemini Intelligence for Android, announced alongside Googlebook, a Gemini-first laptop.
• OpenAI launches the OpenAI Deployment Company, a majority-owned subsidiary with more than $4B in initial capital that embeds forward-deployed engineers inside client organizations, backed by 19 investment and consulting partners led by TPG and paired with a deal to acquire London consultancy Tomoro—a week after Anthropic's own services firm.
• OpenAI Daybreak, a cyber-defense program that gives approved defenders frontier models, the Codex harness, and Codex Security alongside security partners.
• OpenAI Codex in Chrome, an extension that works across background tabs on macOS and Windows.
• OpenAI GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper, new voice reasoning, live translation, and streaming transcription models.
• Google Google Health app, an all-new health app.
• Anthropic Claude for Financial Services, pre-built AI agents for banks targeting pitchbook creation, KYC screening, and month-end close, with new Cowork and Claude Code plugins and native Microsoft 365 integration.
• OpenAI ChatGPT for Excel and Google Sheets, generally available on all plans and powered by GPT-5.5.
• OpenAI GPT-5.5 Instant, the new default model in ChatGPT.
• Anthropic launches AI-native enterprise services firm with Blackstone, Hellman & Friedman, Goldman Sachs, and a consortium including Apollo, General Atlantic, and Sequoia—an Anthropic-engineering-embedded consulting venture targeting mid-market AI deployment at scale.
• OpenAI Advanced Account Security, opt-in protections for accounts at elevated risk of attack.
• Anthropic Claude Security, a public beta that lets Enterprise customers scan their codebases for vulnerabilities and get suggested patches.
• OpenAI models, Codex, and Managed Agents come to AWS, notable given Anthropic's existing AWS relationship—both major frontier labs now distribute through Amazon's cloud.
• OpenAI GPT-5.5, a frontier model for coding, research, computer use, and long-running work, released six weeks after GPT-5.4 to paid ChatGPT and Codex users, with the API following a day later.
• OpenAI Workspace Agents in ChatGPT, automated multi-step workflows across business tools, in research preview for enterprise teams.
• Google Gemini Enterprise Agent Platform, unveiled at Cloud Next '26 for organizations to build, govern, and manage autonomous agents.
• Google eighth-generation TPUs, built for agentic-era workloads.
• OpenAI ChatGPT Images 2.0, with sharper text rendering, multilingual support, and a thinking mode that can search the web and generate multiple images from one prompt.
• Google Deep Research Max, a next-generation Deep Research agent for high-level, autonomous research tasks.
• Anthropic expands Amazon partnership, securing up to 5GW of Trainium capacity and committing $100B in AWS spend over 10 years. Amazon adds a fresh $5B investment, bringing total to $13B.
• OpenAI Codex for (almost) everything, letting Codex use Mac apps, create images, run ongoing tasks, and build memories from screen context with Chronicle.
• Anthropic Claude Design, create prototypes, slides, and one-pagers from natural language—aimed at founders and PMs without a design background.
• OpenAI GPT-Rosalind, a purpose-built model for life sciences research and drug discovery, in research preview for qualified customers via a trusted access program, with early partners including Amgen, Moderna, and Thermo Fisher.
• Anthropic Claude Opus 4.7, a direct upgrade to Opus 4.6 focused on long, hard software engineering, with roughly 3x higher-resolution vision, a new xhigh effort level, and unchanged $5/$25 pricing.
• OpenAI GPT-5.4-Cyber, a cybersecurity-tuned GPT-5.4 for defensive workflows.
• OpenAI $100 ChatGPT Pro tier, a new lower-priced Pro plan.
• Anthropic Claude Cowork, generally available on macOS and Windows with usage analytics and role-based access controls.
• Meta Muse Spark, Meta's first major frontier model from Meta Superintelligence Labs under Alexandr Wang, a direct bid to compete after Llama 4's disappointing reception.
• Anthropic Claude Managed Agents launches in public beta, a fully managed agent harness with secure sandboxing, built-in tools, and server-sent event streaming—running Claude as an autonomous agent through the API.
• Anthropic Claude Mythos Preview and Project Glasswing, a restricted frontier model shared with a small group of partners to find and fix vulnerabilities in critical software ahead of broader Mythos-class releases.
• OpenAI ChatGPT in CarPlay, on iPhones running iOS 26.4 or later.
• Anthropic acquires Coefficient Bio, a stealth AI drug-discovery startup founded eight months earlier by former Genentech Prescient Design researchers, in a stock deal reported at $400 million. Its team of about 10 joins Anthropic's health and life sciences group.
• Google Gemma 4, four open models from edge sizes up to 31B, released under Apache 2.0 for the first time and built for reasoning and agentic workflows. The 31B model debuted third among open models on Arena.
• OpenAI closes at $122B and an $852B valuation, extending participation to retail investors for the first time via bank channels. APIs now processing 15B tokens per minute.
• OpenAI Codex integrations, built-in connections to Slack, Figma, Notion, and Gmail.
• Google Gemini 3.1 Flash Live, its best audio model to date for real-time conversation.
• Google Search Live goes global, reaching more than 200 countries and territories.
• Google switching tools for Gemini, importing memory and chat history from other assistants.
• OpenAI shuts down Sora and pivots to enterprise, cutting "side quests" to focus on coding, agents, and business users. The Sora team redirected toward robotics and world simulation.
• Google Lyria 3 Pro, generating tracks up to three minutes long.
• Anthropic interactive apps on Claude mobile, bringing live charts, diagrams, and app interfaces to iOS and Android.
• OpenAI product discovery in ChatGPT, shopping powered by the Agentic Commerce Protocol with product comparisons and merchant integrations.
• Anthropic computer use in Cowork and Claude Code, a research preview for Pro and Max that lets Claude open files, point, click, and navigate your screen.
• Google full-stack vibe coding in Google AI Studio, turning prompts into working apps with an agent that understands the whole project.
• OpenAI GPT-5.4 mini and nano, smaller, faster GPT-5.4 models for coding, tool use, and sub-agent workloads, with mini available to Free users.
• Anthropic Dispatch, a research preview that lets Pro and Max users direct Cowork from their phone through a persistent thread.
• Anthropic Claude Partner Network, $100M committed for 2026 to train and co-sell with systems integrator partners including Accenture, Deloitte, KPMG, and PwC.
• OpenAI Interactive Learning, new interactive ways to learn math and science in ChatGPT.
• Anthropic interactive charts and diagrams, which Claude now creates inline in its responses.
• Google Ask Maps, a conversational Gemini experience in Google Maps, plus immersive navigation.
• Google Gemini upgrades in Docs, Sheets, Slides, and Drive, for AI Pro and Ultra subscribers.
• Meta acquires Moltbook, the viral social network for AI agents, bringing its founders into Meta Superintelligence Labs.
• OpenAI Codex Security, formerly Aardvark, an application security agent that finds, validates, and patches vulnerabilities, in research preview.
• OpenAI GPT-5.4, its first general-purpose model with native computer use, folding in GPT-5.3-Codex's coding strengths with a 1M-token context window across ChatGPT, the API, and Codex.
• OpenAI ChatGPT for Excel, a beta add-in for building models and running scenarios inside workbooks, with new financial data integrations.
• OpenAI Codex app for Windows, bringing the desktop app to Windows.
• OpenAI GPT-5.3 Instant, for smoother everyday conversations.
• Google Gemini 3.1 Flash-Lite, its fastest, most budget-friendly model yet.
• Anthropic memory for free users, bringing memory from chat history to every Claude user.
• OpenAI raises $110B at a $730B valuation, anchored by $50B from Amazon, $30B from SoftBank, and $30B from NVIDIA—the largest private funding round in history at the time.
• Google Nano Banana 2, combining Nano Banana Pro's image quality with Flash speed.
• Anthropic scheduled tasks in Cowork, for recurring and on-demand tasks.
• Anthropic acquires Vercept, the Allen Institute–rooted startup behind Vy, a cloud computer-use agent that could operate a remote MacBook, and shuts Vy down on March 25.
• Anthropic Cowork enterprise launch with 11 open-source plugins for sales, legal, finance, and healthcare. Framed as "Claude Code for knowledge workers."
• Anthropic Cowork plugin marketplace, with admin controls for Team and Enterprise plans.
• Anthropic Claude Code Security, which scans codebases for vulnerabilities and suggests patches for human review, in a limited research preview.
• Google Gemini 3.1 Pro, more than doubling Gemini 3 Pro's ARC-AGI-2 score to 77.1%, rolling out in the Gemini app, AI Studio, Vertex AI, and Gemini CLI.
• Google Lyria 3 in the Gemini app, for creating custom music from a description.
• Anthropic Claude Sonnet 4.6, full upgrade across coding, computer use, long-context reasoning, and agent planning, with a 1M token context window in beta—becomes the default model on claude.ai and Claude Cowork.
• OpenAI Lockdown Mode and Elevated Risk labels, stricter protection settings for ChatGPT.
• Anthropic raises $30B Series G at a $380B valuation, with Claude Code at $2.5B ARR—more than doubled since January 2026.
• OpenAI GPT-5.3-Codex-Spark, a smaller real-time coding model running on Cerebras hardware for Pro users.
• Google Gemini 3 Deep Think upgrade, improved with scientists and researchers.
• Anthropic self-serve Enterprise plans, so any organization can buy Enterprise without talking to sales.
• OpenAI testing ads in ChatGPT, adding an advertising revenue layer alongside subscriptions.
• OpenAI Frontier, an enterprise platform for building, deploying, and managing AI coworkers with shared business context and explicit permissions, launched with a limited set of customers.
• OpenAI GPT-5.3-Codex, a Codex-native agentic coding model that pairs GPT-5.2-Codex's coding with GPT-5.2's reasoning and runs 25% faster—released the same day as Claude Opus 4.6.
• Anthropic Claude Opus 4.6, the first Opus model with a 1M-token context window (in beta), plus adaptive thinking, effort controls, and agent teams in Claude Code.
• Anthropic Claude for PowerPoint, an add-in for building and editing slides.
• Anthropic Claude will remain ad-free, a direct response to OpenAI's plans to introduce ads to ChatGPT—framing advertising as incompatible with a genuinely helpful AI assistant. Backed by a Super Bowl campaign.
• OpenAI Codex app, a macOS command center for running multiple coding agents in parallel on long-running work, with skills and scheduled automations.
• Google Project Genie, opening its Genie 3 world model to Google AI Ultra subscribers to create and explore interactive worlds.
• Google Chrome auto browse, a Gemini 3–powered agent in Chrome that handles multi-step chores like comparing travel prices or filling a cart within a budget, in preview for AI Pro and Ultra subscribers in the US.
• OpenAI Prism, a free, LaTeX-native cloud workspace with GPT-5.2 built in for scientists to write and collaborate on research papers.
• Google Agentic Vision in Gemini 3 Flash, letting the model inspect images step by step instead of in a single snapshot.
• Anthropic interactive apps in Claude, letting connected tools like Slack, Figma, Asana, and Canva render live, interactive interfaces right inside the conversation.
• Google Personal Intelligence in AI Mode, bringing personalized answers to Search.
• OpenAI ChatGPT Go worldwide, expanding its lower-priced plan globally.
• OpenAI ChatGPT Translate, instant text translation across dozens of languages.
• Google Personal Intelligence, letting the Gemini app securely connect Gmail, Google Photos, YouTube, and Search to give answers tailored to you.
• Anthropic Labs, a new team for incubating experimental products, with Instagram co-founder Mike Krieger joining to build alongside Ben Mann.
• Google Veo 3.1 updates, adding more consistency and control to Ingredients to Video.
• Anthropic Cowork research preview, agentic desktop computing for general knowledge workers, extending the Claude Code model beyond developers. Enterprise software stocks dropped on the day of launch.
• Anthropic Claude for Healthcare, HIPAA-ready tools spanning providers, payers, and patients—prior authorizations, care coordination, and regulatory submissions on the enterprise side, with Pro and Max subscribers able to connect health records, labs, and Apple Health or Android Health Connect data for plain-language summaries. Announced five days after ChatGPT Health, timed to the J.P. Morgan Healthcare Conference.
• Google Universal Commerce Protocol, an open standard for agentic shopping unveiled at NRF 2026 so AI agents can check out with retailers directly from AI Mode and Gemini.
• Google Gmail enters the Gemini era, with upgraded Help me write, AI Overviews, and personalized suggested replies.
• OpenAI ChatGPT Health, a dedicated space where users link patient portals, Apple Health, and wellness apps to ask questions grounded in their own labs, visit summaries, and insurance documents—stored separately from other chats and excluded from model training, built with 260+ physicians across 60 countries. OpenAI says 230 million people already ask health questions weekly.
• Meta acquires Manus, maker of a general-purpose AI agent, in a deal reported at more than $2B, with plans to fold its agents into Meta AI and Meta's enterprise offerings.
• OpenAI personality controls, letting you adjust ChatGPT's warmth, enthusiasm, and emoji use.
• OpenAI GPT-5.2-Codex, an agentic coding model for complex real-world software engineering.
• Anthropic Claude in Chrome for all paid plans, a beta with Claude Code integration and workflow recording, as Skills become easier to deploy across organizations.
• Google AI video verification in the Gemini app, letting people check whether a video was made with Google AI.
• Google Gemini 3 Flash, frontier-level intelligence at Flash speed, becoming the default model in the Gemini app and AI Mode in Search.
• OpenAI app submissions for ChatGPT, letting developers publish apps to a new in-ChatGPT directory.
• OpenAI new ChatGPT Images, built on GPT Image 1.5 for more precise edits and faster generation.
• Google Gemini 2.5 Flash Native Audio update, for more natural voice interactions and complex workflows.
• OpenAI GPT-5.2, released days after Google's Gemini 3 and Anthropic's Claude Opus 4.5—an unprecedented compression of frontier model releases within a single month.
• Google Gemini Deep Research agent for developers, available through the Interactions API.
• Google GenTabs, a Google Labs browsing experiment that turns open tabs into interactive apps.
• Google studio-quality virtual try-on, a more personalized way for US shoppers to try on clothes.
• Anthropic Accenture multi-year partnership, Accenture Anthropic Business Group formed, 30,000 Accenture professionals trained on Claude, joint CIO offering for enterprise AI adoption built around Claude Code.
• OpenAI, Anthropic, and Block launch the Agentic AI Foundation (AAIF) under the Linux Foundation, donating MCP, AGENTS.md, and goose as open standards for agent interoperability. AWS, Google, Microsoft, and Bloomberg join as platinum members.
• Anthropic Claude Code reaches $1B ARR six months after public launch, reportedly faster than any enterprise software product in history—announced alongside the acquisition of Bun, the JavaScript runtime used to scale Claude Code's infrastructure.
• Anthropic Snowflake $200M partnership, Claude available to 12,600 Snowflake enterprise customers across Bedrock, Vertex, and Azure for governed AI on structured and unstructured data.
• OpenAI ChatGPT Shopping Research, personalized buyer's guides pulling from across the web, timed for Black Friday. Part of a broader push to make ChatGPT an e-commerce destination.
• Anthropic Claude Opus 4.5, setting new marks in coding, agents, and computer use while cutting Opus pricing to $5/$25 per million tokens.
• Anthropic Claude for Excel beta, for Max, Team, and Enterprise users, as context compaction enables nearly endless conversations.
• Google Nano Banana Pro (Gemini 3 Pro Image), a viral image generation model rolling out across Gemini, NotebookLM, Google Slides, Ads, and Workspace.
• OpenAI GPT-5.1-Codex-Max, the new default Codex model, built for long-running agentic coding.
• Meta SAM 3, a Segment Anything model that finds, segments, and tracks every instance of a concept named in a text prompt across images and video.
• Anthropic Microsoft, NVIDIA, and Anthropic announce strategic partnerships, scaling Claude on Azure with a $30B compute commitment and making Claude the only frontier model available on all three major clouds. NVIDIA and Microsoft commit up to $10B in investment.
• Google Gemini 3, its most intelligent model, shipped on day one in the Gemini app and AI Mode in Search.
• Google Google Antigravity, an agent-first development platform where agents plan, write, and verify code across the editor, terminal, and browser.
• Google Gemini 3 in Search, shipping in AI Mode on launch day.
• Google agentic travel planning in AI Mode, helping build itineraries and book trips.
• Google WeatherNext 2, generating forecasts 8x faster at up to one-hour resolution.
• OpenAI group chats in ChatGPT, piloted in a few regions before expanding to all logged-in users on November 20.
• Google SIMA 2, an agent that plays, reasons, and learns with you in virtual 3D worlds.
• Google agentic checkout, headlining new holiday AI shopping features in Search and Gemini.
• OpenAI GPT-5.1, Instant and Thinking upgrades that are warmer and more conversational and adapt how long they think to the task.
• OpenAI Sora app for Android, bringing the video app beyond iOS.
• Google Gemini in Google Maps, for hands-free, conversational driving help.
• OpenAI Aardvark, an agentic security researcher that scans code for vulnerabilities and proposes patches, in private beta.
• OpenAI company knowledge in ChatGPT, answering from a workspace's connected tools with citations for Business, Enterprise, and Edu.
• Anthropic memory for Pro and Max, extending Claude's memory beyond Team and Enterprise.
• Google Samsung Galaxy XR, the first Android XR headset, built around Gemini.
• OpenAI ChatGPT Atlas, a new AI browser with agentic web capabilities.
• Google vibe coding in Google AI Studio, turning a described idea into a working AI app.
• Anthropic Claude Code on the web, letting Pro and Max users hand coding tasks to Claude running in Anthropic-managed cloud sandboxes from the browser.
• Anthropic Claude for Life Sciences, connecting Claude to Benchling, PubMed, 10x Genomics, and other scientific platforms, with agent skills for research workflows.
• Anthropic Agent Skills, folders of instructions, scripts, and resources Claude loads when relevant, working across the Claude apps, Claude Code, and the API.
• Anthropic Claude Haiku 4.5, its small model matching Sonnet 4 on coding, computer use, and agent tasks at a third of the cost.
• Google Veo 3.1, adding richer audio and more narrative control to its Flow filmmaking tool.
• Google Gemini Enterprise, chat with your company's documents, data and apps as well as build and deploy AI agents, all grounded in your information and context.
• Anthropic IBM strategic partnership, Claude added to IBM's developer IDE for enterprise software workflows, with a joint guide for deploying enterprise-grade AI agents.
• Google Gemini 2.5 Computer Use, a model that lets developers' agents operate browser and mobile interfaces by clicking, typing, and scrolling, available in the Gemini API.
• OpenAI Apps in ChatGPT, chat with third-party services right in ChatGPT.
• Anthropic Deloitte expanded alliance, Claude available to 470,000 Deloitte people globally, Claude Center of Excellence, and a certification program for 15,000 practitioners—Anthropic's largest enterprise AI deployment at the time.
• OpenAI AgentKit, a DevDay toolkit for building, deploying, and optimizing agents, with a visual Agent Builder, a Connector Registry, and ChatKit.
• OpenAI Sora 2 and GPT-5 Pro in the API, announced at DevDay 2025.
• Google Gemini for Home, replacing Google Assistant on smart speakers and displays alongside a redesigned Google Home app.
• OpenAI Sora 2 and the Sora app, a video-and-audio model with better physics and control, launched alongside an invite-only social iOS app built around cameos of yourself and friends.
• Google visual search in AI Mode, letting people search and get inspired with images.
• OpenAI Instant Checkout for Merchants, and the underlying protocol developed jointly with Stripe is open source.
• Anthropic Claude Sonnet 4.5, new benchmark records in coding, reasoning, and computer use. Accompanied by the launch of the Claude Agent SDK for building capable agents.
• OpenAI parental controls in ChatGPT, linking parent and teen accounts with quiet hours, content settings, and safety notifications.
• James Campbell on Twitter: OpenAI Pulse, a daily briefing of everything going on in your life...yet another step towards ChatGPT becoming a true personal assistant.
• Meta Vibes, a short-form feed of AI-generated videos in the Meta AI app where people can create, remix, and share clips to Instagram and Facebook.
• Google Gemini Robotics 1.5, bringing AI agents into the physical world.
• Deedy Das on Twitter: "I read through every single job posting on OpenAI's career page. Here's everything they're working on and who they compete with.
• Anthropic's new Keep thinking ad.
• Google Gemini in Chrome, an AI browsing assistant that works across your open tabs, rolling out to desktop users in the US.
• Google shareable Gems, letting people share custom Gems like Google Drive files.
• Meta Ray-Ban Display glasses, AI glasses with an in-lens display for notifications, navigation, translation, and Live AI—paired with the Meta Neural Band for gesture control. Available at $799 from September 30.
• OpenAI GPT-5-Codex, a version of GPT-5 optimized for agentic software engineering that adapts how long it thinks to the task.
• Anthropic memory for Team plans, plus incognito chats for all users.
• OpenAI MCP tools in ChatGPT, available through developer mode.
• Anthropic file creation in Claude, letting Claude create and edit Excel spreadsheets, PowerPoint decks, documents, and PDFs in the app.
• OpenAI Projects for Free users, plus conversation branching.
• OpenAI gpt-realtime, its most advanced speech-to-speech model, as the Realtime API becomes generally available.
• Anthropic code execution tool, letting Claude run Python in a sandbox through the API.
• Google Nano Banana (Gemini 2.5 Flash Image), an image generation and editing model that went viral and topped image-editing leaderboards.
• Anthropic Claude for Chrome, a research-preview browser extension that lets Claude navigate sites, click, and fill forms, piloted with 1,000 Max users.
• Google live translation in Google Translate, plus AI-powered language practice.
• Google agentic features in AI Mode, letting Search do more on your behalf, like finding restaurant reservations.
• Anthropic premium seats with Claude Code, for Team and Enterprise plans.
• OpenAI ChatGPT Go, a lower-priced plan launched first in India.
• Anthropic offers Claude to all three branches of the U.S. government for $1—the broadest AI offer to federal workers to date, covering executive, legislative, and judicial branches. One week after a similar but executive-branch-only offer from OpenAI.
• Anthropic 1M-token context for Claude Sonnet 4, in beta on the API.
• Anthropic search across past conversations, for Max, Team, and Enterprise users.
• OpenAI GPT-5, a unified model with built-in thinking that becomes the default for ChatGPT users.
• Google Jules, its asynchronous coding agent, leaves beta and becomes available to everyone.
• Anthropic automated security reviews in Claude Code, via a /security-review command and GitHub Actions.
• Google Guided Learning, an interactive study partner in the Gemini app.
• Others in the industry are witnessing the AI lab productization trend play out. Here is some strategic advice from a VC to founders on how to potentially survive this new paradigm.
• OpenAI gpt-oss-120b and gpt-oss-20b, its first open-weight language models since GPT-2, released under Apache 2.0.
• Anthropic Claude Opus 4.1, an upgrade to Opus 4 for agentic tasks, real-world coding, and reasoning.
• Google Genie 3, a general-purpose world model that generates interactive environments, shown as a research preview.
• Google Deep Think in the Gemini app, for Google AI Ultra subscribers.
Menu |
About The Author |
More AI Writing |
|---|---|---|
|
|
Joy Larkin is a technologist in Silicon Valley. She likes robots and is excited for Superintelligence.
|
Frontier AI Lab Tracker
|