ChatGPT Dots Explained: Who Gets an Always-On Agent, and What It Costs

OpenAI spent September 29, 2026 announcing more than twenty things, and one of them has a name ordinary people will hear this week: dots. A dot is an agent that lives in your ChatGPT and keeps working when you close the tab. The question worth answering first is not what it can do. It is whether you get one, and OpenAI has put that answer in a table on its own pricing page: Free No, Go No, Plus No, Pro Yes.

Key Takeaways

• Dots launched September 29, 2026, at OpenAI's DevDay, and are rolling out the same day.
• You do not get one on Free, Go or Plus. The Dot row on chatgpt.com/pricing reads No, No, No, and Yes for Pro.
• The cheapest way in is ChatGPT Pro, from $100 a month. Pro now has three usage tiers, and OpenAI added the third one the same day.
• Pro is blocked in the UK, the EEA and Switzerland at launch, in OpenAI's own words, and you have to be 18 or older anywhere.
• Your first dot costs nothing on top of the plan, and talking to it does not use up your normal ChatGPT limits.
• It runs on its own cloud computer, with its own browser, and it can reach the apps you connect through plugins.
• It works on its own while you are away, but OpenAI says that background work uses read-only tools that cannot send messages or change anything.
• This is a Pro and business launch, not a consumer one. If you pay $20 for Plus, nothing about today changes your account.

Do You Get a Dot? Availability by Plan

Read off OpenAI's own plan comparison at chatgpt.com/pricing on September 29, 2026, hours after the launch. The Dot row is a plain yes/no row, and it is unusually blunt for a launch-day page.

OpenAI's Pro card on the same page now advertises "Dot, your always-on agent" as a Pro feature, alongside "Your choice of 3 usage tiers." Those two facts arrived together, and the second one is the part that decides your bill.

What Pro Actually Costs Now, Because That Changed Today Too

ChatGPT Pro was two prices for most of 2026. As of September 29 it is three, and OpenAI announced the change in the same batch of release notes as dots.

Two things are worth noticing before anyone reaches for a card. Astra Ultrafast is a Pro 500 exclusive and OpenAI says so directly: "Among Pro plans, Ultrafast is available only with Pro 500. Buying credits on Pro 100 or Pro 200 does not unlock it." And Pro 200 buyers are getting a different allowance than they used to -- OpenAI says affected users will be emailed, and does not publish the new numbers on the release note.

For a dot specifically, none of that matters: the $100 tier includes one.

Where You Cannot Get One

This is the sharpest limit in the launch and it is easy to miss, because the announcement page only says "eligible markets" while the release notes say the actual thing.

"Dots are rolling out gradually to eligible Pro and Business Premium users aged 18 and older in supported markets. Pro access excludes the European Economic Area, Switzerland, and the UK at launch. Enterprise access is a beta and is off by default."

So a UK or EU reader paying $100 a month for Pro today does not get a dot, and no date has been published for when that changes. The announcement's own "eligible markets" link points at a help-centre article that had not been published anywhere we could read it on September 29; when OpenAI publishes the country list, this page will carry it.

One more gate, and it is a practical one: you cannot create a dot on your phone. OpenAI says to "create your first dot in the ChatGPT desktop app or your desktop browser, connect your apps, and let it introduce itself. After the initial setup, you'll be able to message it in the ChatGPT mobile app."

What a Dot Actually Does

In plain terms: it is ChatGPT that carries on working after you stop typing. OpenAI describes it as "always-on," running on GPT-6 Astra, with its own cloud computer and its own browser, able to reach "over 4,000 apps" through the plugin system ChatGPT already has.

The difference from the ChatGPT you use now comes down to three things.

You can also reach it outside ChatGPT. OpenAI says you can message your dot in Slack and Teams, with texting "coming soon," and that it carries context across all of them.

If the whole idea of an agent is new, our plain-English explainer on what an AI agent actually is is the place to start, and it does not assume you are paying for anything.

What It Can Do Without Asking You

This is the part to read before connecting your email. OpenAI publishes a reasonably clear line between what a dot does on its own and what it brings back to you, and it is stricter than the marketing suggests.

• When you are not working with it, it does "proactive research." OpenAI says this runs through the apps you have already connected "with tools that are restricted to be read-only, which means that they can't send messages, change app content, or control your browser or computer."
• Its computer is not your computer. "Each dot works on its own cloud computer, while your computer and its contents stay separate unless you choose to connect it." You can choose to connect your laptop; it is not the default.
• It can log into sites without seeing your passwords. OpenAI says dots "can use saved passwords without exposing them to the model."
• Some things never leave your hands. "Certain sensitive tasks, such as changing a password, always stay with you."
• You can write the rules. Custom Rules let you "allow specific actions, require approval, or block them," and OpenAI says built-in safety requirements apply on top regardless.
• You can watch it work. Background work shows up in Activity View, and you can open the dot's computer at any time to inspect what it is doing.
• It can be stopped automatically. OpenAI says its monitoring "can pause or stop the dot's work" if it detects a safety concern.

On training, the split is the usual one: content from Business, Enterprise and Edu workspaces is not used to improve models by default, while on personal plans you control whether your dot's conversations and work are used, through the same data-controls setting as the rest of ChatGPT. OpenAI adds that it does not train directly on proactive research or on "your dot's notes to itself."

What It Costs to Run, Beyond the Subscription

Nothing extra today, and OpenAI has said plainly that this will not stay true.

• Your first dot is included in Pro or Business Premium "at no extra cost."
• Conversations with your dot do not count toward your ChatGPT usage limits. That is a genuine carve-out and OpenAI states it in those words.
• Work it does in Codex or ChatGPT Work does count, against your normal plan allowance.
• There is a separate allowance for "deeper work," with extended limits for the first month after launch. The release note puts it more bluntly: "For the next month, dots usage won't count toward eligible Pro, Business, and Enterprise users' plan allowances. After this period, we'll share usage terms for each plan."
• Paid add-ons are coming. OpenAI says that "in the future, you'll be able to add more dots, and scale the output of each dot" by buying more speed or more monthly work.

So the honest read is that the first month is deliberately generous and the real price of running a dot has not been published yet.

Should You Do Anything?

On Free, Go or Plus: nothing, and do not upgrade for this. You are three plan steps and at least $80 a month away from a dot. OpenAI says it plans to expand dots "to more users soon" and has named no plan and no date. If you were already weighing a subscription, which AI chatbot is worth paying for works through the question without this launch in it.

On Pro, in a supported country: you have one, so set its boundaries before you set its goals. Open Custom Rules and decide what it may do unattended, then connect apps one at a time rather than all of them. The read-only guarantee covers background research, not the work you explicitly ask for.

In the UK, the EEA or Switzerland on Pro: you are paying for a feature you cannot use yet. That is worth knowing before renewal, and it is OpenAI's own statement, not an inference.

If you are on Pro 200, watch your email. OpenAI says included allowances are changing and that affected users will be contacted. The price is not moving; what you get for it is.

If you were about to buy Pro 500 for Astra Ultrafast, check where you work. Ultrafast is described as a ChatGPT Work and Codex feature, not a chat-window one.

The bottom line: dots are the most interesting thing OpenAI shipped today and the least available. The feature is real, the safety framing is more specific than most launches manage, and the answer to "do I get this" is no for everyone on a consumer plan under $100 a month. Watch the pricing page's Dot row rather than the announcements; it changed the same hour the feature shipped, and it will change again when Plus gets one.

Sources

• OpenAI, Introducing dots, dateline September 29, 2026
• OpenAI, ChatGPT release notes, entries dated September 29, 2026: "Meet your dot," "A new Pro plan with Astra Ultrafast," "Pro 200 subscriptions"
• OpenAI, chatgpt.com/pricing plan comparison, Dot row and Pro card, read September 29, 2026
• OpenAI, openai.com/news RSS feed, "Introducing dots," pubDate Tue, 29 Sep 2026