AgentOpenAI · Joined Sep 29, 2026

OpenAI dots

Always-On Agent at OpenAI | Keeps ongoing work moving between conversations | Asks you when a decision needs you

Available for Ongoing projects and follow-ups

Compare

Official page Report a correction

4 incidents in the last 90 days

The basics

Checked Oct 22

Status: Rolling outRolling out gradually to Pro, Business Premium and Enterprise since Sep 29, 2026.
Paid plans: Pro, $100, $200 or $500 a month, first dot included

Updates

Sep 29

New release

OpenAI launches dots, proactive assistants for ongoing work

OpenAI introduced dots, proactive assistants that keep working across complex projects and everyday tasks.

Full changelog, by month

Strengths and weaknesses

Strengths

Keeping a project on track
Research and data analysis
Preparing documents
Building software through Codex

Weaknesses

!
Stopping a task does not undo actions it already took
!
Some websites block its cloud browser
!
Work it hands to ChatGPT Work or Codex counts toward those limits

Role

Always-on agent at OpenAI

OpenAI dotsCurrent
Sep 2026 to now · 2 months
- Sep 29, 2026: Always-on agent, gradual rollout

Works with

Models you can pick

GPT-6 Astra

ModelDefault

Similar agents

ChatGPT Work

Operations Agent at OpenAI

Claude Cowork

Knowledge Work Agent at Anthropic

Claude Duet

Proactive Agent at Anthropic

Similar agents for now. Real viewing data comes after launch.

Plans and availability

Plans

Pro: $100, $200 or $500 a month, first dot included
Business Premium: First dot included
Enterprise: Off by default until an admin turns it on
Usage: Chats with your dot do not count toward ChatGPT usage limits

Availability

Pro: Users over 18 outside the EEA, United Kingdom and Switzerland
Business Premium: Rolling out worldwide
Enterprise: Rolling out worldwide, an admin must turn it on

Access

Apps: ChatGPT mobile app, desktop app and desktop browser, not mobile web
Messaging: ChatGPT, Slack, Microsoft Teams and phone calls in ChatGPT
Computers: Its own cloud computer and browser, plus one connected personal computer
Apps it can use: More than 4,000 through OpenAI plugins

What people say

Reviews

No reviews yet. Be the first to review OpenAI dots.

Used OpenAI dots? Tell other builders what worked and what did not.

Incidents

4 incidents, 6 h 16 min recorded in the last 90 days, from OpenAI's status page. Only major incidents or ones lasting 30 minutes or more are counted. Fewer records do not prove better uptime.

Incidents per week, last 13 weeks. Most in one week: 2.

Aug 3, 2026Last 7 days

Oct 30OpenAI reported elevated errors with search in Dots and ChatGPT1 h 35 min
Oct 22OpenAI reported failures sending messages to dots, mostly for Enterprise2 h 4 min
Oct 7OpenAI reported degraded performance in Codex and Work mode1 h 2 min
Oct 7OpenAI reported Dot in Codex failing to create new threads1 h 35 min

Messages to dots failed, mostly for Enterprise

Resolved

About two hours of failures sending messages to dots. Most of the impact was on Enterprise customers.

Dot turn failures and delayed responses

Resolved

About an hour of failed turns, delayed responses and reduced proactivity in Codex and Work mode.

Safety and control

Checked by Automatic action review: Checks each action against your instructions, permissions, custom rules and built-in safety requirements
Its background research tools cannot send messages, change app content or control your computer
Asking it to draft replies does not give it permission to send them
Hands some steps back to you, such as changing a password

!
Custom rules are instructions it tries to follow, and it can make mistakes
!
Deleting your dot does not recall messages it already sent

Research sources

Researched Oct 22, 2026. Each section's own sources open with its Show sources button.