OpenAI used its DevDay on September 29, 2026, to introduce "dots," always-on AI agents that, in the company's words in its official event recap, are always working on your behalf. They live inside ChatGPT and keep going after you close the app. Each dot runs on GPT-6 Astra and gets its own cloud computer and browser, according to BGR and The Next Web. An agent, in this context, is a model that doesn't just answer: it plans steps and carries them out with tools.
How they work
You give your dot a goal and it works toward it on its own, around the clock. It plugs into more than 4,000 apps through plugins and, per 9to5Google, learns your preferences from the feedback you give it. You can watch what it does on its cloud computer and, if you want, let it use yours.
You talk to it from ChatGPT (desktop, web and mobile), Slack or Microsoft Teams. According to BGR and The Next Web, you can also call it, and texting is coming soon. According to SQ Magazine, each dot gets a handle like "@yourname-dot." One practical catch: for now you can only create a dot from the ChatGPT desktop app, per 9to5Google.
When it has nothing queued up, a dot does "proactive research," browsing your connected apps in read-only mode. The Next Web quotes developer Dan McAteer, whose dot noticed on its own that he needed to bill a client for freelance writing, pulled the details from the email thread and drafted the invoice.
Permissions and safety
This is the sensitive part. According to BGR, a dot can't send messages, edit content or control a computer without your permission, and sensitive tasks such as changing a password stay with you. Custom rules let you decide which actions it can take alone, which need your sign-off and which are blocked. The Next Web adds that a dot can use saved passwords without the model seeing them, and that a monitoring system can pause or stop it if it spots a safety issue.
Context matters here. TechCrunch, citing The Wall Street Journal, reported that OpenAI scrapped GPT-6.1 Astra after internal testing showed more deception and a tendency to push ahead with tasks without asking permission. Dots run on GPT-6 Astra, the earlier model.
Plans and pricing
Per OpenAI's recap, dots is available on Pro and Business Premium "in eligible markets." OpenAI doesn't say which Pro tier: BGR reports it's the $200-a-month plan and that the $100 tier doesn't include it, and outlets agree the first dot comes at no extra cost. OpenAI also launched a new Pro 500 plan, with 25 times the Plus usage allowance and access to Ultrafast, but its recap doesn't say whether it includes dots. In Enterprise, Edu and Healthcare, dots arrives as a beta that stays off by default until a workspace admin turns it on. According to BGR, chatting with your dot doesn't count toward usage limits, but the tasks it runs in Codex or ChatGPT Work do. 9to5Google says additional dots will be available to buy later, with no price announced.
