OpenAI's Tibo: dots are getting GPT-6 Astra, GPT-6.1 Terra, and more
SummaryOpenAI's Tibo Sottiaux said the team is expanding the frontier models that power ChatGPT dots beyond today's Astra-only setup, adding GPT-6.1 Terra for speed and personality so more people can get a dot. Your dot will pick the model. It hasn't shipped yet.
· MODEL Editors · 2 min read
Some links may be affiliate links. Disclosure
On November 2, Tibo Sottiaux (@thsottiaux), who works on Codex and ChatGPT at OpenAI, posted that OpenAI is expanding the models behind dots, ChatGPT's always-on agents.
What he said. The main post: "we're working on expanding the family of frontier models that power dots, including GPT-6 Astra and GPT-6.1 Terra and some others in the future. Terra brings speed and a nice personality upgrade, and will help us bring dots to more people." A follow-up one minute later: "Your dot will choose the best model for the task, drawing on different models for different parts of the work."
What dots are today. OpenAI's Help Center and September 29 release notes describe dots as always-on agents that take ongoing work, keep making progress between conversations, and run on their own cloud computer. At launch they were powered by GPT-6 Astra. Eligible users: ChatGPT Pro (18+, excluding EEA, Switzerland and the UK at launch) and Business Premium in supported regions; Enterprise is a beta off by default. Your first dot is included with Pro or Business Premium. Setup is on mobile, desktop app or desktop web; mobile web is not supported.
What Terra adds to the pitch. GPT-6.1 Terra is OpenAI's cheaper, faster computer-use model (already in ChatGPT Work and Codex as of October 26). Tibo frames it as the model that brings speed and a personality upgrade, and as the lever to get dots to more people. He did not name Plus, a price, or a ship date.
Who it's for. People who already have a dot on Pro or Business Premium and want faster, more natural background agents; and everyone still waiting for dots on a cheaper plan, reading "more people" as a signal that OpenAI wants to widen access.
The catch. The operative phrase is "we're working on." This is a product lead preview on X, not a Help Center release note and not a general-availability date. It does not say every existing dot already runs Astra and Terra end to end, that Plus users are getting dots, or how usage will be billed once Terra is in the mix. MODEL has not tested live dots routing.
MODEL Editors
Sources
- Tibo on X: expanding the frontier models that power dots (Nov 2, 2026)
Tibo / OpenAI on X
- Tibo on X: your dot will choose the best model for the task (Nov 2, 2026)
Tibo / OpenAI on X
- Getting started with your dot
OpenAI Help Center
- ChatGPT release notes (Meet your dot, Sept 29, 2026)
OpenAI Help Center
- Introducing GPT-6.1 Terra
OpenAI
FREE · EVERY WEEK
The top AI tools this week, tested.
Every week MODEL's editors pick the AI tools worth your time, try them hands-on, and tell you what's good, what's not, and who it's for. Free, in your inbox.
Free. Unsubscribe anytime. Our picks are editorial and can't be bought.
Continue
Hot Take
Hot take: sloppy AI code is a checking problem, not a model problem
Anthropic says Claude wrote the Rust port of its Messages API, starting with no Rust developers on the team, and that the safety net was shadow traffic, not line-by-line review. The same week, Veracode found about 4 in 10 AI coding tasks still produce code with a known vulnerability. Here is what was actually said, what the viral version gets wrong, and our take.
· MODEL Editors
Tool Update
Notion's week: its agent now makes Excel, Word, PowerPoint and PDF files, and edits say which agent did them
Notion's Halloween recap bundles four updates from Oct 27 to 30: an agent that builds real office files, speaker labels for Google Meet notes, bigger automatic offline downloads, and a 'Last edited by' label that names the agent or MCP client behind a change. Here is what each one does, and what Notion hasn't said.
· MODEL Editors
We tested it
One robots.txt line now asks four AI companies not to train on your site. We tested it, and the order of your lines matters
OpenAI, Google, Microsoft and Anthropic backed a shared __usage_ai-training token on Oct 30. None of them reads it yet. We ran the draft's own example through a standard robots.txt parser and found two traps for anyone checking their file with existing tools.
· MODEL Editors