OpenAI introduces dots, always-on AI agents in ChatGPT
OpenAI added new AI agents called dots to ChatGPT. They keep working in the background even when no one is talking to them.
Each dot runs on GPT-6 Astra. It has its own cloud computer and connects to more than 4,000 apps.
Users can reach their dot through ChatGPT, Slack, Teams or a voice call, with texting coming soon. And it doesn’t just wait for commands, it reaches out first. It comes to users with progress updates, questions and decisions that need their approval. Users can open its computer anytime to watch what it’s doing.
In one of OpenAI’s examples, a developer’s dot reads customer feedback, then prepares and tests small fixes on its own. All the developer has to do is review and approve.
Users stay in control. When no one’s talking to it, a dot works read-only and can’t send messages. Sensitive actions go through an automated check, and users decide which ones need their approval. OpenAI’s own warning: dots can still make mistakes. For now it’s only on Pro and Business Premium.
Sources
- OpenAI, “Introducing dots”, (openai.com)
- 9to5Google, “OpenAI launches Dots, new ‘always-on agents’ you can assign tasks to”, (9to5google.com)
- The Decoder, “OpenAI launches always-on Dots agents to rival Meta’s Muse”, (the-decoder.com)
- Tom's Guide, “OpenAI announces ChatGPT Dots — here’s what the new AI agents can do”, (tomsguide.com)
- The Next Web, “OpenAI launches dots, always-on AI agents with their own cloud computers”, (thenextweb.com)
- The New Stack, “OpenAI just launched Dots. Here’s why they matter for developers.”, (thenewstack.io)
About this story
This story was posted on Instagram by @jarrus.tech on Sept. 30, 2026.
Spotted an error in this story? [email protected] · Instagram
Short link: thejarrus.com/en/openai-dots
This story in Turkish: OpenAI, ChatGPT için 7/24 çalışan yapay zeka ajanları dots’u tanıttı
On the same topic
Robot tests: even the best model completed only 19% of 84 robot tasks, and 63 tasks were solved by no model
In RobotWorld, a benchmark released October 7, the top model, GPT-6 Astra, completed only 16 of 84 simulated robot tasks; no model solved 63 of the tasks.
OpenAI’s image model posts a near-perfect score on text-dense images
On UltraText Bench, a dense-text test led by Westlake University, OpenAI’s GPT Image 2 ranked first of 24 configurations with 99.35 out of 100.
HackerRank’s AI interviewer opens to all customers after more than 500,000 job interviews
HackerRank made its AI interviewer Chakra generally available to customers on October 5, after a roughly six-month beta with more than 500,000 interviews.