Latest AI News
Latest AI news, in plain words
What changed in Claude, ChatGPT, Gemini, Midjourney, Higgsfield, Seedance and other big AI tools, newest first, each with who it matters for and a link to the source.
Last checked: Nov 5, 2026
ChatGPT
Released Effective Rolling out now, no fixed date
ChatGPT can now make animated GIFs from a text prompt or a photo you upload, and you can change the speed or edit single frames. It runs on the new Images 2.6 model, which also makes pictures faster in ChatGPT and Codex. For everyone.Source ↗Higgsfield
Released Effective Nov 3, 2026
Higgsfield launched Shorts, which makes vertical 9:16 reels hosted by your own AI character. You give a topic and a short reference video for the movement, and it writes the script for you, for clips from 10 seconds to 2 minutes. For creators.Source ↗Claude
Released Effective Nov 3, 2026
Text written by Claude Mythos 5, Claude Opus 4.7 and Claude Sonnet 4.6 through Anthropic's developer platform now carries a hidden watermark, so it can later be recognized as AI-made. With Sonnet 4.6, changing settings like temperature may stop the mark from being added. For builders.Source ↗Claude
Released Effective Nov 2, 2026
Developers can now try message threads (beta). Anthropic's servers keep the conversation so far, so each new request only needs to send the new messages. This makes chat apps on Claude simpler to build. For builders.Source ↗ChatGPT
Released Effective Rolling out now, no fixed date
OpenAI is merging the Chat and Work tabs in the ChatGPT desktop app into one place, so you can go from an idea to getting it done without switching tabs. Codex stays separate but shares the same usage allowance. Some users get it first. For desktop app users.Source ↗Higgsfield
Released Effective Nov 2, 2026
Higgsfield launched Mangaka, a tool that turns an idea or a script into manga, comics or webtoons, so you can make an illustrated story without drawing it yourself. For Higgsfield users.Source ↗Claude
Released Effective Oct 30, 2026
The paid plans have new names but the same prices: "Max 5x" is now Max at $100 a month, and "Max 20x" is now Ultra at $200 a month. Verified Ultra accounts no longer hit the 5-hour limit, but the weekly limit still applies. For paid plans.Source ↗Claude
Released Effective Oct 30, 2026
Claude Code can now automatically shorten (compact) long conversations that have been sitting idle, including in VS Code and the Claude desktop app. Long coding sessions stay cheaper and keep running smoothly. For builders.Source ↗ChatGPT
Released Effective Oct 30, 2026
On iPhone and Android you can now stay signed in to two ChatGPT accounts and switch between them without logging out. Each account keeps its own chats, memory and settings. For everyone on mobile.Source ↗Higgsfield
Released Effective Oct 30, 2026
Higgsfield added Intone, which makes videos of talking characters whose lips match the words and whose voice stays the same from clip to clip. Useful for presenters, explainers and AI influencers. For Higgsfield users.Source ↗Gemini
Released Effective Oct 29, 2026
Gemini can now create and edit documents, slides, spreadsheets and PDFs right in the chat. Most formats work on the web only; PDFs also work on mobile. For everyone.Source ↗Claude
Released Effective Oct 23, 2026
The single Claude app, where chat and Cowork are combined, is now fully available to Enterprise companies, so their staff can start bigger, multi-step tasks from any conversation. For Enterprise plans.Source ↗Seedance
Released Effective Oct 23, 2026
ByteDance's cloud arm released Seedance 2.5 4K for film and TV work: 4K video, 10-bit color, more natural faces and small movements, and cheap 480p drafts before the final render. For selected business customers (trial).Source ↗Claude
Released Effective Oct 22, 2026
Anthropic gave Agent Skills, the open format for packaging instructions that teach an AI how to do a job, to the Agentic AI Foundation under the Linux Foundation. A neutral group now looks after it, so other AI tools can use the same skills. For builders.Source ↗Claude
Released Effective Oct 21, 2026
Claude Duet is a new app in early preview. It checks your email and calendar, drafts replies and prepares research on its own, and asks you before it sends, pays for or deletes anything. US only, by invitation from a waitlist. For Pro and Max plans.Source ↗Higgsfield
Released Effective Oct 21, 2026
Kling 4.0, a new video model that can also edit clips, is now inside Higgsfield, so you can try one of the newest video models without leaving the site. For Higgsfield users.Source ↗ChatGPT
Released Effective Oct 20, 2026
Locked chats let you hide private conversations from your history and search. They open only with a 4-digit PIN or your phone's Face ID or fingerprint. For Free, Go, Plus and Pro.Source ↗ChatGPT
Released Effective Oct 16, 2026
ChatGPT Slides builds presentations from a chat, a file or an outline. You can edit them with others and export to PowerPoint, PDF or Google Slides. For eligible Pro, Business and Enterprise plans.Source ↗ChatGPT
Released Effective Oct 15, 2026
Voice mode now runs on a new model, GPT-Live-1.1, with more natural accents and new voices for several languages. Find them in Settings, then Voice. For everyone.Source ↗Other tools
Released Effective Oct 15, 2026
Kling 4.0: Kuaishou's new video model is open to everyone. It makes 3 to 30 second clips at up to 4K with stereo sound, and you can steer it with up to 15 reference pictures, videos or voices. For everyone on Kling.Source ↗ChatGPT
Released Effective Oct 14, 2026
The older GPT-5.5 model was removed from ChatGPT, so you pick a newer model from the menu instead. For everyone.Source ↗Gemini
Released Effective Oct 13, 2026
Gemini 4 Argon, Google's newest top model, is now in the Gemini app. It handles complex tasks better, gives more direct answers, and can keep working on tasks while you are away. For AI Pro and Ultra.Source ↗Gemini
Released Effective Oct 13, 2026
Skills began rolling out in the Gemini app: saved instructions you can reuse and run automatically. They replace Gems. For everyone with a personal Google or Workspace account.Source ↗Gemini
Released Effective Oct 13, 2026
A new Customize menu puts response style, skills and connected apps in one place. Gemini now connects to 40+ apps such as Instacart, OpenTable, Adobe and Shopify. For everyone (most apps US English only).Source ↗Gemini
Released Effective Oct 9, 2026
The free Gemini app now offers only the lighter Flash-Lite model. AI Plus keeps Flash but loses Pro, while AI Pro and Ultra keep all models. For free and AI Plus users.Source ↗Midjourney
Released Effective Oct 8, 2026
Midjourney is testing Thinking Mode: click "Rerun (Thinking)" and it compares the picture with your prompt, spots mistakes like wrong objects, text or hands, and makes it again. For Alpha site testers.Source ↗Claude
Released Effective Oct 7, 2026
Anthropic released Claude Haiku 5.5, its smallest and fastest model. Anthropic says it costs about 75% less to run than Haiku 4.5 and is much more capable, good for quick, high-volume jobs like summaries and sorting. For everyone.Source ↗Claude
Released Effective Oct 7, 2026
Max, Ultra and Team subscriptions now include free API credits each month to run your own apps and agents on the Claude Platform: $100 for Max and $200 for Ultra. For paid plans.Source ↗ChatGPT
Released Effective Oct 7, 2026
GPT-6 with "Intelligent UI" adds interactive charts, buttons and calculators right inside answers, and can start replying while it is still thinking. You can turn the visuals down if you want fewer. For everyone (Free and Go from October 8).Source ↗Midjourney
Released Effective Oct 7, 2026
Midjourney added shared folders: invite people to add pictures or just look, or share by link. It makes teamwork and client reviews easier. For Alpha site users.Source ↗Gemini
Released Effective Oct 6, 2026
Nano Banana 2.1, Google's faster picture maker and editor, is rolling out in the Gemini app. It keeps faces and logos more consistent, writes text more sharply and can make pictures up to 4K. For Gemini app users.Source ↗Claude
Released Effective Oct 6, 2026
Anthropic expanded its startup program: eligible young companies get a free year of Claude Team with up to five seats, plus $1,000 in API credits. For startups.Source ↗Higgsfield
Released Effective Oct 2, 2026
Higgsfield launched AI Influencer: design your own AI character, then put that same character into videos, keeping one face across all your content. For Higgsfield users.Source ↗ChatGPT
Released Effective Sep 29, 2026
OpenAI launched Dots, always-on AI helpers. Each one has its own cloud computer, connects to your apps and keeps working toward your goals after you close the chat. For Pro and Business Premium now; Plus, Go and Free later.Source ↗ChatGPT
Released Effective Dec 11, 2026
Custom GPTs stop working and are replaced by plugins. If you made a GPT, use the built-in "Migrate to plugin" button before then. Your old chats stay. For everyone who uses or made GPTs.Source ↗Claude
Released Effective Sep 28, 2026
Anthropic released Claude Sonnet 5.5, a faster and cheaper partner to Opus 5.5, built for everyday tasks and polished documents, slides and spreadsheets. It is the fastest Sonnet so far. For everyone.Source ↗Gemini
Released Effective Nov 17, 2026
Google starts turning your saved Gems into skills automatically. Your instructions and files carry over, and you run a skill by typing "/" in the chat. For everyone who uses Gems.Source ↗Claude
Released Effective Sep 25, 2026
A new developer portal lets people submit plugins to the Claude directory, follow the review and see usage numbers after launch, which should mean more ready-made add-ons inside Claude. For builders.Source ↗Midjourney
Released Effective Sep 24, 2026
Midjourney's editor got better at inpainting and outpainting: it now changes only the area you select, and quality holds over repeated edits. The Styles sidebar can preview your prompt in different styles, and --tile patterns no longer show seams. For subscribers.Source ↗Midjourney
Released Effective Sep 23, 2026
Korean is now available on midjourney.com, the Alpha site lets you save your favorite settings as defaults, and the Create feed has a new full-width layout. For subscribers.Source ↗Other tools
Released Effective Sep 23, 2026
Google Vids: Google's newest video model, Gemini Omni 1.1 Flash, now makes free 1080p AI videos in Google Vids on a computer, with scene extension and exact length control. For anyone with a Google account; higher limits need a paid plan.Source ↗Claude
Released Effective Sep 22, 2026
Anthropic released Claude Opus 5.5, the first model in the 5.5 family: strong at coding and complex work, answers more than 30% faster, and costs about 40% less to run than Opus 5. Paid plans also got higher 5-hour limits. For everyone.Source ↗Seedance
Released Effective Sep 22, 2026
Higgsfield added Seedance 2.5 Draft Mode: make cheap 480p test videos first, then render only the one you like in 1080p. This saves credits while you experiment. For Higgsfield users.Source ↗Claude
Released Effective Sep 16, 2026
Chat and Cowork are now one Claude, so you can start bigger tasks from any chat without switching modes. Claude can also make designs, slide decks (Claude Slides) and documents (Claude Docs) in any conversation, including Claude Code. Pro and Max first, other plans following.Source ↗Gemini
Released Effective Sep 10, 2026
Gemini now has a Windows desktop app that opens with Alt + Space. It can work with your Gmail and Drive files and make Nano Banana pictures. For Windows 10 or later.Source ↗ChatGPT
Released Effective Sep 8, 2026
ChatGPT Images 2.5 makes sharper pictures up to 50% faster and edits only the part you ask for. You can also type "@Sketch" to draw a rough idea and turn it into a picture. For everyone.Source ↗ChatGPT
Released Effective Sep 3, 2026
GPT-6 Astra, OpenAI's most capable model yet, arrived in ChatGPT Work and Codex for harder tasks. For Pro, Enterprise and Business Premium first, then Plus and Business.Source ↗Midjourney
Released Effective Sep 2, 2026
Midjourney's V8.2 editor now opens right inside the picture viewer on the Alpha site: edit with plain instructions, add up to 4 reference pictures, and see every edit in one place. A "Change Style" button is being tested. For Alpha site users.Source ↗Claude
Released Effective Sep 1, 2026
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1, its most advanced models for coding and knowledge work. Fable 5.1 is available to everyone on the API; Mythos 5.1 is limited to a special program called Project Glasswing. For builders and paid plans.Source ↗Higgsfield
Released Effective Sep 1, 2026
Higgsfield launched Genjutsu, which swaps objects, places, products or characters in an existing video and can copy motion from one clip to another, so you can remake a video without shooting it again. For Higgsfield users.Source ↗Claude
Released Effective Nov 30, 2026
The older Claude Sonnet 4.5 model stops working on the Claude API, and fast mode for Claude Opus 4.8 is removed. Apps that use either one need to switch to Sonnet 5.5 or Opus 5.5. For builders.Source ↗Gemini
Released Effective Sep 2026
Gemini Live voice got a new model, Gemini 3.8 Live, which also lets you get things done by voice in Gmail, Docs and Keep. For everyone.Source ↗Midjourney
Released Effective Aug 27, 2026
Midjourney released its first V8.2 Edit model for everyone to test: change pictures by typing instructions, combine up to 4 reference pictures (this replaces omni-reference), and fix or extend parts of a picture. For subscribers.Source ↗Claude
Released Effective Aug 25, 2026
Memory now works in both chat and Cowork. You can see and edit what Claude remembers about you, and sensitive subjects such as health or beliefs are left out unless you turn on a setting. For everyone.Source ↗Gemini
Released Effective Aug 19, 2026
Gemini added a student hub and study notebooks on mobile. Eligible college students can get a free year of AI Pro in the US, or AI Plus in 140+ countries, if they claim it by December 31, 2026. For students.Source ↗ChatGPT
Released Effective Aug 18, 2026
ChatGPT for Teens is a version for younger users, with stronger safety protections and study tools built in. For teens and their parents.Source ↗Higgsfield
Released Effective Aug 12, 2026
Higgsfield released Cinema Studio 4.0, with 30-second generations, up to 50 reference pictures, and new camera and color controls, for more film-like control over each scene. For Higgsfield users.Source ↗Seedance
Released Effective Aug 7, 2026
Seedance 2.5 arrived in CapCut on computers, with AI Video, AI Edit and AI Extend, so you can make, fix and lengthen clips in the same editor, paying with credits by the second. For CapCut Pro and Ultra members.Source ↗Seedance
Released Effective Aug 7, 2026
Seedance 2.5 opened to developers through BytePlus ModelArk (international) and Volcano Ark (China), so other apps and websites can build it in and it will show up in more tools. For developers.Source ↗Seedance
Released Effective Aug 3, 2026
Seedance 2.5 came to Higgsfield: clips up to 30 seconds long with matching sound in a single generation, so you stitch fewer short clips together. For Higgsfield users.Source ↗Seedance
Released Effective Aug 3, 2026
ByteDance announced Seedance 2.5, launched worldwide on Dreamina on July 31: 30-second videos (up to 3 minutes in Long Video Mode), up to 50 reference files, and fixing one spot in a video without redoing the whole thing. For paid Dreamina plans.Source ↗Gemini
Released Effective Aug 2026
Gemini Omni 1.1 Flash, Google's video maker, arrived in the Gemini app. It can make clips longer, fill in the motion between a first and last frame, and upscale to 4K. For Gemini app users.Source ↗Other tools
Released Effective Sep 24, 2026
Sora: OpenAI shut Sora down in two steps. The app and website closed in April, and the developer API and Sora 2 models switched off in September. If a tutorial tells you to use Sora, choose another video tool. For everyone.Source ↗