OpenAI Dots: Always-On AI Agents on GPT-6 Astra (DevDay 2026)
At DevDay on September 29, 2026, OpenAI introduced Dots — hosted, always-on agents powered by GPT-6 Astra, each with its own cloud computer and ~4,000 app connectors. What a Dot does, who gets it today, the usage fine print, and the new Pro tiers, with OpenAI and The Register linked throughout.
OpenAI puts an always-on agent in ChatGPT
On September 29, 2026, at DevDay in San Francisco, OpenAI introduced Dots: persistent, hosted AI agents that keep working on long-running tasks. According to The Register's DevDay report, Dots run on GPT-6 Astra, each get their own cloud computer, can work toward goals 24/7, retain context over time, and can use some 4,000 apps via connectors. OpenAI's own announcement is Introducing dots.
This page is the long-form briefing behind the BSH Technologies Instagram carousel.
What a Dot actually is
In the apps, a Dot shows up as a friendly animated shape. Underneath, as The Register puts it, Dots are persistent compute tasks that consume tokens: a model, a set of tools, and a long-running job. Sam Altman described his Dot reviewing what came in overnight and pinging him in the morning with anything urgent before his day starts.
- Model: GPT-6 Astra.
- Compute: each Dot gets its own cloud computer and can pursue goals around the clock.
- Reach: roughly 4,000 apps via connectors.
- Memory: retains context over time and learns how you like to work.
- Collaboration: Dots can work inside ChatGPT Space pages — you can tag a Dot in a comment and it gets to work.
Who gets Dots today
- Pro and Business Premium plans in eligible markets, with "an allowance for deeper work, with generous limits for the first month after launch" (OpenAI spokesperson to The Register).
- Enterprise: beta available pending admin approval.
The usage fine print
Conversations with your Dot do not count toward ChatGPT usage limits. But when you ask your Dot to start or manage tasks in Codex or ChatGPT Work, those tasks count toward your limits as usual. OpenAI also says that in future you'll be able to pay a flat monthly fee to add more Dots or increase their speed or amount of work.
The new Pro tiers around it
- New $500/month Pro: 25× Plus usage, plus Ultrafast — up to 8× faster token generation with GPT-6 Astra in ChatGPT Work and Codex (an API version runs at 6× standard speed).
- $200/month Pro reopens (suspended around September 10) with reduced limits: 10× Plus usage, down from 20×.
- GPT-6.1 Sol was also announced, pitched as near-Astra intelligence at one-fifth the price.
Why it matters
- Agents leave the chat window. A Dot is a background worker, not a reply box — the design question becomes what you let it do unattended.
- Connectors are the attack surface and the value. 4,000 apps means real leverage and real permission scoping work for IT and security teams.
- Cost is usage-shaped. Chat is free of limits, but the work a Dot kicks off in Codex or ChatGPT Work is not — budget for tasks, not messages.
Primary sources
- OpenAI — Introducing dots (Sep 29, 2026)
- The Register — OpenAI tries disarming AI angst with cute graphics and always-on agents (Sep 29, 2026)
How BSH can help
At BSH Technologies we help teams decide where always-on agents belong — which connectors to allow, what a Dot may do unattended, and how to budget usage. If you are piloting Dots or building your own background agents, our Thrissur engineers can map permissions, guardrails, and cost before you roll it out.
Frequently asked questions
What are OpenAI Dots?
Dots are hosted, always-on AI agents OpenAI introduced at DevDay on September 29, 2026. They run on GPT-6 Astra, each get their own cloud computer, can work toward goals 24/7, retain context, and use roughly 4,000 apps via connectors.
Who can use Dots today?
Pro and Business Premium subscribers in eligible markets, per an OpenAI spokesperson quoted by The Register. Enterprise customers can try a beta pending admin approval.
Do Dots count against my ChatGPT usage limits?
Conversations with your Dot do not. Tasks your Dot starts or manages in Codex or ChatGPT Work count toward your usage limits as usual.
Related Topics
From the blog
View all posts
Nvidia Open Agent Safety Platform: OpenShell + Sentry on BlueField-4
Jensen Huang introduced Nvidia’s Open Agent Safety Platform on Sep 28, 2026 — OpenShell software paired with Sentry on BlueField-4 DPUs that Nvidia says can quarantine a rogue agent breakout in milliseconds. Supporters listed include Anthropic, Arm, Microsoft, Oracle, and SpaceX; OpenAI is not listed. Primary TechCrunch reporting linked throughout.
Oracle Force Majeure on New Mexico Stargate Data Center (Project Jupiter)
Oracle sent a force majeure notice to the developer of Project Jupiter, a Stargate AI data center campus in New Mexico — not exiting as main tenant, but able to delay payments if the 2028 online target slips. Oracle says it does not expect a delay. Primary TechCrunch / Bloomberg reporting linked throughout.