Dots, OpenAI's always-on agents

2026-09-29 · Publisher : OpenAI · Domain : Artificial Intelligence

OpenAI launches dots, always-on agents powered by GPT-6 Astra: each has its own cloud computer, connects to the chosen apps and moves projects forward proactively, under rules set by the user.

OpenAI launches dots, always-on agents powered by GPT-6 Astra. Each one has its own computer in the cloud, connects to the apps the user chooses and moves projects forward proactively. ### In detail - A dot receives a goal and autonomy rules, then works continuously, including on several projects in parallel. - It has its own identity for access, its own browser, and can write and test code; users can look at its computer at any time. - It can be reached in ChatGPT (web, mobile, desktop), by SMS, in Slack and Teams, or by voice call, keeping context across channels. ### Specifications - More than 4,000 connectable apps through the plugin ecosystem, according to OpenAI. - Proactive research in the background is limited to read-only tools. - Custom rules (allow, require approval, block), automatic verification of sensitive actions and an Activity view. - Availability: Pro and Business Premium in eligible markets; beta for Enterprise, Edu and Healthcare when enabled by the admin. - First dot included in the plan; conversations do not count against ChatGPT limits, tasks started in Codex or ChatGPT Work do. ### Why it matters OpenAI moves from an assistant that answers to an agent that takes on ongoing responsibility. Specialized dots, with their own identity inside companies, are in pilot, and an integration with Microsoft Agent 365 is in progress. ### Key takeaways Dots can still make mistakes: OpenAI recommends checking any work with significant consequences. The list of eligible markets remains to be checked for Europe.

Tags : OpenAI, Agent, ChatGPT, GPT-6, Automatisation