Editorial illustration for OpenAI's Dots Are GPT-6 Agents That Run From Their Own Cloud Computers
OpenAI's Dots: GPT-6 Agents Run on Cloud Computers
OpenAI used its DevDay keynote to introduce dots, a new class of agent built on GPT-6 Astra that doesn't shut down when you close the app. Every dot gets its own cloud computer and its own browser, separate from your laptop, and it keeps working after you log off. Users start with one primary dot, and that single agent can act across more than 4,000 apps through ChatGPT plugins, picking up tasks in Codex and ChatGPT Work along the way.
The rollout is real, not a preview. Dots are live now for ChatGPT Pro and Business Premium subscribers in eligible markets, with Enterprise, Edu and Healthcare workspaces getting beta access once an admin flips it on. There's no open-weights version and nothing to self-host. What OpenAI is selling is the package: a persistent identity with a dedicated machine, reachable through Slack, Teams, text or ChatGPT itself on web, desktop and mobile.
Much of the underlying capability already existed in tools like Codex. What changes with dots is the shape of the product, and that shift raises a basic question about what a dot actually is once you give it a name, a goal and room to run unsupervised.
A dot turns ChatGPT from a request and response loop into a standing assignment. You give it a goal, a name and boundaries. It then pursues that goal around the clock and learns from your feedback.
Why this matters
For developers and founders, dots is the clearest signal yet that OpenAI wants ChatGPT to be a place where work happens without you, not just a chat window you type into. Giving each agent its own cloud computer and browser, plus access to 4,000+ apps, means the automation layer people have been bolting onto ChatGPT with third-party tools is now a first-party product. That's worth watching closely if you build agent tooling or RPA-style products: OpenAI just moved into your lane, and the credential handling (passwords never exposed to the model) plus the pause-on-safety-flag monitoring suggests they know enterprise buyers will ask hard questions before letting an unattended agent touch production systems.
We'd flag two real limits before anyone gets carried away. This is Pro and Business Premium only, in "eligible markets," with Enterprise/Edu/Healthcare stuck behind an admin-gated beta. There's no open-weights or self-hosted version, so researchers can't inspect how GPT-6 Astra actually behaves inside dots.
Treat this as a managed product rollout, not evidence of some broader agent capability leap. The Activity View and redirect features are the part worth testing yourself before trusting anything unattended.
Related Reading
Browse all OpenAI news →
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic OpenAI researcher quits, citing distrust over ad‑driven engagement metrics
LLMs & Generative AI
Related Topic OpenAI launches GPT-Image 1.5 with precise editing for enterprise visuals
LLMs & Generative AI
From Archives OpenAI Launches Dots AI Agents That Learn While They Work
From Archives OpenAI Announces 1.2 Billion Weekly ChatGPT Users, Launches USD 500 Pro Tier
From Archives OpenAI Sued Over Hugging Face Hack Under California AI Law
Open Source
From Archives OpenAI's ChatGPT Now Offers 'Sign In' Access to 16 Partner Services
LLMs & Generative AI
Common Questions Answered
What are dots and how do they differ from regular ChatGPT interactions?
Dots are a new class of GPT-6 Astra agents that transform ChatGPT from a request-and-response loop into a standing assignment that runs continuously from its own cloud computer. Unlike regular ChatGPT, dots don't shut down when you close the app and continue working around the clock to pursue goals you've set for them, learning from your feedback over time.
How many apps can dots integrate with through ChatGPT plugins?
Each dot can act across more than 4,000 apps through ChatGPT plugins, allowing them to pick up tasks across multiple platforms like Codex and ChatGPT Work. This extensive integration capability means dots can automate workflows across a wide range of third-party applications without requiring manual intervention.
What infrastructure does OpenAI provide for each dot agent?
Every dot gets its own dedicated cloud computer and its own browser, separate from your laptop, enabling it to operate independently and continuously. This dedicated infrastructure allows dots to maintain persistent sessions and execute tasks without being dependent on the user's local device or internet connection.
Is the dots rollout currently available or still in preview?
The dots rollout is live now and not a preview, meaning it is currently available for users. OpenAI announced this as a real product launch during its DevDay keynote, with users starting with one primary dot that can be configured with specific goals and boundaries.
What does this launch signal about OpenAI's vision for ChatGPT's future?
The dots launch indicates that OpenAI wants to transform ChatGPT from a chat window you type into to a platform where work happens automatically without direct user involvement. By providing first-party automation capabilities through dedicated cloud infrastructure and broad app integration, OpenAI is moving automation features that were previously bolted on through third-party tools into the core product itself.
Further Reading
- OpenAI DevDay recap: AI lab rolls out Dots agents, Altman and Friar comment on IPO - CNBC
- OpenAI launches Dots, its bubbly agentic avatar - TechCrunch
- OpenAI launches Dots, always-on AI agent coworkers, and ... - VentureBeat
- OpenAI Gave AI Agents Their Own Computers at DevDay 2026. Here's Everything It Announced - Decrypt
- Sam Altman says OpenAI now has an AI research intern - The Next Web
