OpenAI is starting to roll out ChatGPT for Teens globally today (August 18, full availability in Australia expected September 8) to Free and paid personal accounts, switching it on automatically when the system estimates someone is under 18 or they state their age is between 13 and 17 - ChatGPT for Teens bundles teen onboarding, Study Mode, quizzes and learning visualizations, responsible homework reminders that redirect shortcut attempts to Study Mode, Study Hours, break reminders and sensitive image upload reminders - Protections are on by default and based on the Under-18 Principles in the Model Spec, covering self-harm, violence, eating disorders, dangerous activities and explicit sexual or graphic content, and parents with linked accounts can set quiet hours, manage selected settings and get safety notifications (now including eating disorder notifications) without reading conversations OpenAI also updated the Model Spec today with clearer principles on relational interactions for teens (ChatGPT should not use romantic language, encourage emotional dependence or imply it has feelings or consciousness), is adding under-18 evaluations to system cards, and announced a partnership with CodeAI on AI education for students and educators

Hey @qatarairways, I chose the Vegetarian Jain Meal (VJML) today because it matches my dietary restrictions, though not for religious reasons. I noticed that the kathi roll with squash and chickpeas listed onion as an ingredient. Just a heads-up in case this is an error, since onion would normally conflict with a Jain meal and could cause issues for passengers relying on the VJML designation.

BREAKING: The US and China have agreed to extend their trade deal until January 10th, 2027. “I don’t know whether a bigger deal can be done. I don’t know whether we will just roll the current deal,” US Treasury Secretary Bessent said.

OpenAI released GPT-6 Sol and Luna with API prices cut in half, and Anthropic released Claude Opus 5.5, the first Claude 5.5 model, at roughly Fable 5.1 level for 40% less than Opus 5, with Sonnet 5.5 and Haiku 5.5 announced for the coming weeks - GPT-6 Sol and Luna are trained like Astra (still the top model), cost 50% less than GPT-5.6 promotional pricing ($2/$10 and $0.10/$0.50 per million tokens), get 90% off cached input reads plus cache-safe effort and tool toggles, write shorter and clearer, show fewer misleading claims about coding work, and roll out gradually today in ChatGPT Work (not yet in Chat), Codex and the API for paid users, Luna in the desktop app for Free and Go - Opus 5.5 costs $4/$20 per million tokens with $0.20 cache reads, $2/$10 batch and $8/$40 fast mode, outputs over 30% faster than Opus 5, writes clearer, thinks more per turn, reads charts and screenshots better, keeps zero data retention and adds EU AI Act watermarking - Anthropic is raising the 5-hour usage limits on Pro, Max, Team and seat-based Enterprise plans and giving subscribers a rate limit reset they can save and use whenever they choose, similar to the banked resets OpenAI previously added to Codex - Opus 5.5 scores best on Anthropic's behavioral audit and gets Fable 5.1-class safeguards - offensive cyber work falls back to Opus 4.8, dual-use biology and a small set of frontier LLM development tasks (like kernel development for certain ML accelerators) fall back to Opus 5

OpenAI introduced today at DevDay 2026 more than 20 updates across ChatGPT, Codex, models and the API, including dots (always-on agents), GPT-6.1 Sol (close to GPT-6 Astra at a fifth of the price), Ultrafast (a speed tier with up to 8x faster token generation), ChatGPT Space (shared pages you edit together with people and agents) and Pro 500 (a new $500/month plan) Models - GPT-6.1 Sol comes close to GPT-6 Astra on agentic coding, computer use and professional work at one-fifth of Astra's token prices ($2 input, $10 output, $0.10 cached input per 1M tokens), available in the API as gpt-6.1-sol and in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu, not yet in Chat - Ultrafast generates tokens up to 8x faster (300 tokens per second) in Codex and up to 6x faster in the API, with GPT-6 Astra Ultrafast available today in the API, ChatGPT Work and Codex on Pro 500 and Enterprise, and GPT-6.1 Sol Ultrafast coming in the next days - Ultrafast uses included plan usage at 8x the Standard rate and purchased credits or Enterprise pay-as-you-go at 6x, is off by default for Enterprise workspaces and isn't available to workspaces that need inference residency outside the US - Private Intelligence adds per-project policies with Zero Data Retention, Private Safety Processing and validated external storage Dots - Dots run on GPT-6 Astra, have their own cloud computer and browser, connect to over 4,000 apps through plugins, learn from feedback and keep working between conversations, and you can message or call them in ChatGPT, Slack or Microsoft Teams - Dots can run several tasks in parallel with background agents, decide themselves when to pause and wake up, and create Work or Codex tasks, including in Codex cloud environments - Dots research your connected apps in the background with read-only tools, and actions that affect your accounts or share information go through auto-review, custom rules and built-in safety checks - You can open your dot's cloud browser at any time to take over a step and hand control back, dots can sign in to websites with saved passwords without exposing them to the model, and you can connect your computer through your dot's profile to give it access to local files, code and apps - Texting your dot is coming next, Pro users can join a waitlist for iMessage or RCS on Android, and at launch a dot can use your personal email but can't get its own standalone email address - Your first dot is included in your Pro (Pro 100 too), Business Premium or Enterprise plan at no extra cost, conversations with it don't count toward your ChatGPT usage limits, tasks it starts or manages in ChatGPT Work or Codex count toward those limits as usual - Dots roll out gradually to Pro users outside the EEA, UK and Switzerland, Business Premium worldwide and Enterprise as an off-by-default beta, you create them on desktop only (mobile app support follows), and OpenAI plans teams of dots per user, specialist dots for companies as enterprise pilots and an integration into Microsoft Agent 365 Pro plans - Pro 500 is a new $500/month plan with the most included usage and the only Pro tier with Ultrafast (buying credits on Pro 100 or 200 doesn't unlock it), and Pro 200 is available again for new subscribers with a lower usage allowance, with existing subscribers keeping their previous allowance through Oct 29, 2026 ChatGPT Space and collaboration - ChatGPT Space replaces Library for Pro, Business and Enterprise with pages you edit in real time with team members, comment on, tag @ ChatGPT or @ dot in and fill with generated text, images, interactive visualizations and prompt blocks, with collaborative slides, spreadsheets and mobile editing coming later - Each collaborator uses their own ChatGPT, sharing a page doesn't share your private chats or memory, Projects stay separate from Space, Enterprise sharing is an opt-in preview, and pages that refresh from connected tools (Keep Updated) aren't available at launch - The Meetings plugin (beta, macOS desktop app, Pro and Business, Enterprise coming soon) takes notes from what's said in the meeting, saves summaries with action items to Space and deletes the audio once the notes are ready - @ ChatGPT works in Slack and Teams with a deployment's connected tools, and Enterprise gets Team Tasks that run scheduled or event-triggered work through a team's service account, workspace connections for company-managed accounts and Agent Security policies for tools, files and network access - Shareable profiles (private by default) show your Sites and plugins, Sites get in-place editing and scheduled updates, and synced Work tasks continue across desktop, web and mobile Codex - Codex gets reusable cloud environments where tasks run in isolated workspaces and continue with your computer asleep (Plus, Pro, Business, Healthcare, Education and Enterprise), an updated Codex CLI and a Code Review plugin for pull requests - Codex Security Cloud (research preview) scans connected GitHub repos on demand, on a schedule or on new commits, checks findings, removes duplicates and prepares patches as draft pull requests, with access to Daybreak Blue models only inside Security Cloud API and developers - New Decisions API (constrained decision-making on Luna with image inputs in under a few hundred milliseconds), the Agents API with computer use and Bedrock Managed Agents on AWS running OpenAI models - Plugin extensions let developers add sidebar apps, conversation panels, file viewers and editors, and forms to ChatGPT, plugin creation, submission and discovery got updated, and Sites can now host plugins - MCP Events (needs MCP 2.0, webhook delivery only) let plugins start automations when something happens in a connected app, available on all plans - Sign in with ChatGPT lets you sign in to third-party apps with your name, email and profile picture (no conversations or memory shared), and eligible Plus and Pro users can use their ChatGPT plan allowance in participating apps with an app-specific limit (limited trial with selected partners) - A Browser Annotation API lets websites choose what people can select and attach context before sending feedback to ChatGPT Enterprise and other - OpenAI Marketplace lets eligible enterprise customers put part of their existing OpenAI commitment toward 32 partner tools, including Adobe, Figma, Sierra, Salesforce, ServiceNow, Harvey, Palo Alto Networks, CrowdStrike and Baseten - OpenAI now counts 1.2B weekly users, its team rewrote ChatGPT Web on top of the desktop app codebase in 5 weeks with Astra so future launches ship on web and desktop on the same day, and reports Responses API token traffic up 100x year over year at 99.9% uptime

The @SpaceX Dragon lifted off at 6:33 a.m. EDT today. It will arrive at 11 a.m. on Thursday, Oct. 15, to resupply the Expedition 75 crew with over 6,300 pounds of science, supplies, and roll-out solar arrays. More...