OpenAI introduced Dots at DevDay on September 29: agents intended to take responsibility for ongoing work, rather than finish a single chat response. Each dot has its own cloud computer and browser and can keep working between conversations.
The launch summary asks users to give a dot a goal and define what it may do independently. It can work across connected apps, bring back results for review and ask when a decision needs the user’s judgment.
OpenAI’s Dots overview identifies GPT-6 Astra as the underlying model and describes tasks such as research, data analysis, document preparation and software development. Relevant past conversations and preferences provide context for that work.
Cloud work can continue while the user’s own computer is off. The launch described a gradual rollout to eligible accounts, not immediate availability for everyone.