Breaking

OpenAI Introduces Dots as Always-On AI Agents for ChatGPT Users

September 30, 2026 at 23:44 EDT

AI Agents
Foundation Models

OpenAI introduced Dots, a new class of always-on AI agents in ChatGPT designed to keep working on user-directed projects between conversations. The agents, powered by GPT-6 Astra, have their own cloud computer and browser, can use connected apps and tools, and are being rolled out gradually to eligible accounts.

September 29, 2026 · OpenAI

Dots make ChatGPT persistent

OpenAI introduced always-on agents that keep working between conversations, use a cloud computer and ask for approval when user judgment is needed.

The first rollout starts with one primary assistant

primary dot per person at launch

The numeric story is deliberately small: OpenAI is starting users with a single assistant, then asking them to define permissions, connected apps and review rules before handing over recurring work.

GPT-6 Astra

Core model behind Dots

Cloud

Own computer and browser

Review

Approvals for sensitive actions

How a dot is meant to work

Remember

Use context, memory and saved notes.

→

Work

Research, create files and use tools.

→

Check

Ask for approval when judgment matters.

Why teams may try it

Developers and researchers can hand off recurring follow-up: code changes, analysis updates, document preparation and project tracking.

Why caution remains

Persistent app access raises the stakes for permissions, instruction-following and approval design when an agent acts over time.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.