OpenAI introduces dots for ongoing work
OpenAI introduced dots at DevDay as agents that take on an ongoing responsibility. They keep making progress between conversations and return with results or a decision that needs a person’s judgment.
A responsibility that continues between chats
The DevDay release notes describe a dot working across connected apps on its own computer. The person supplies a goal and defines what the agent can do independently. Access is rolling out gradually to eligible accounts.
Work continues between conversations
- Ongoing work
- Person sets the goal and independent-action scope
- Connected apps
- Dot works on its own computer
- Local Work resources
- Connected computer needs to be online
- Access
- Gradual rollout to eligible accounts
View data
| Responsibility | Required boundary |
|---|---|
| Ongoing work | Person sets the goal and independent-action scope |
| Connected apps | Dot works on its own computer |
| Local Work resources | Connected computer needs to be online |
| Access | Gradual rollout to eligible accounts |
OpenAI · published 2026-09-29. Source-bound illustration, not a performance benchmark.
Download imageThe ongoing model changes the handoff: a person need not prompt every individual step, but still needs a clear review boundary. The notes also distinguish synced Work tasks that use an online connected computer for local files and tools. A persistent agent does not imply every local resource remains available when its computer is offline.
Original sources
- OpenAI: original announcementlearn.chatgpt.com
Checked 9 Oct 2026 · A manually curated edition. Availability may change; company performance claims are not Trion test results. Editorial method.