What Are OpenAI Dots? The Always-On AI Agents Explained

Here’s a question that would’ve sounded like science fiction two years ago: what does your AI assistant do when you’re not talking to it? As of September 29, 2026, OpenAI has an answer. So what are OpenAI Dots? They’re always-on AI agents inside ChatGPT that keep working toward your goals between conversations — even after you close the tab.

Announced at DevDay 2026, Dots are the biggest shift in how ChatGPT works since ChatGPT itself. Let’s break down what they actually are, how they work, who can get one, and whether you should care.

What Are OpenAI Dots, in Plain English?

A Dot is a persistent, named AI agent attached to your ChatGPT account. Regular ChatGPT works one turn at a time: you ask, it answers, it stops. A Dot keeps going.

Think of the difference this way. Asking ChatGPT to analyze this month’s customer feedback has a clear beginning and end. Giving a Dot the ongoing responsibility of keeping track of customer feedback is different — new feedback keeps arriving, so the work never really finishes. That’s the mental shift: you’re not giving the AI a bigger task, you’re giving it an ongoing responsibility.

Each Dot runs on GPT-6 Astra (OpenAI’s newest flagship model), gets its own cloud computer and browser, and can connect to more than 4,000 apps through OpenAI’s plugin ecosystem. You can give your Dot a name, an avatar, even a handle like @yourname-dot. You talk to it through ChatGPT, Slack, Microsoft Teams, or voice — and a task you start in Slack can continue in ChatGPT. It keeps context across channels.

How Dots Actually Work

Under the hood, three things make Dots different from the chatbot you’re used to:

1. They have their own computer. Each Dot runs on a separate cloud machine where it browses the web and runs tools — without touching your laptop. That means it can do background research, draft documents, and prepare changes while you’re doing literally anything else.

2. They remember. Dots maintain long-term context about your projects and preferences across sessions. You don’t re-explain the situation every time you open a chat.

3. They check in. A Dot can contact you with progress updates, questions, or decisions that need human judgment. It can also be redirected mid-task — you stay in charge of the direction.

OpenAI’s own examples: a Dot could keep a sales proposal updated as customer requirements change, investigate a newly reported bug, or prepare content the moment new source material arrives.

Dots vs Regular ChatGPT: What’s Really Different?

The honest answer? Most people don’t need always-on agents for most things. If your AI use is “write me an email” or “explain this concept,” Dots change nothing for you.

Where Dots matter is multi-step work with waiting in the middle — the kind of work where a regular chatbot forces you to be the project manager. Research that takes hours. Proposals that evolve over weeks. Monitoring tasks that never end. Dots absorb the coordination overhead: the checking back, the remembering, the “oh right, I was supposed to follow up on that.”

One important clarification: Dots are not the old “ChatGPT Agent” mode. OpenAI says that mode is discontinued — a Dot is a persistent assistant that can start work tasks, not a one-off agent session.

Who Can Get a Dot — and What Does It Cost?

Here’s the part everyone asks about. At launch:

ChatGPT Pro users (Pro $100, $200, and $500 tiers) get their first Dot included at no extra cost — but Pro users in the EEA, UK, and Switzerland are excluded for now.
Business Premium users get it worldwide.
Enterprise workspaces get a beta once an admin enables it.
Plus and Free users? Not at launch. OpenAI says it plans to expand access, with no date given.

Your first Dot is included in the plan. OpenAI says you’ll eventually be able to add more Dots or more capacity, but hasn’t announced pricing. Conversations with your Dot don’t count against your ChatGPT limits — but tasks it kicks off still count as usual.

Real Things People Are Using Dots For

Beyond OpenAI’s examples, the pattern is clear: Dots shine at ongoing responsibilities, not one-off tasks. A developer has one watching customer feedback, flagging bugs, and drafting fixes for review. A product manager uses one to keep launch materials current as requirements shift. Sales teams track changing customer requirements and keep proposals fresh. Researchers re-run analyses when new data lands.

The common thread? Each of these used to require someone to remember to check, re-read the context, and do the work. The Dot holds the context permanently, so the “remember to check” part disappears.

Safety: What Dots Can and Can’t Do Alone

This is where you should pay attention. An always-on agent with app access is powerful — and OpenAI has built real guardrails:

Background work is read-only. Between conversations, a Dot researches but can’t send messages or change files on its own.
Consequential actions need your approval. Sending an email, editing a file, making a purchase — you review these first.
Some things stay human-only. Password changes, for example, are always yours to do.
Memory is partly opaque. You can’t inspect or edit individual Dot memories line by line — and OpenAI’s own testing found that errors (“drift”) grow as task chains get longer.

The practical takeaway from OpenAI’s own system card: Dots are genuinely useful for well-scoped, long-running work — and genuinely risky for anything where a mistake can’t be undone. Stopping a Dot doesn’t reverse actions it already completed. Review anything consequential. That’s not fine print; it’s the whole game with persistent agents.

FAQ

What are OpenAI Dots?

OpenAI Dots are always-on AI agents in ChatGPT, announced at DevDay on September 29, 2026. Each runs on GPT-6 Astra with its own cloud computer, connects to 4,000+ apps, and keeps working between conversations.

How much do OpenAI Dots cost?

Your first Dot is included with ChatGPT Pro ($100/$200/$500 per month) or Business Premium at no extra cost. Pricing for additional Dots hasn’t been announced.

Can I get a Dot on ChatGPT Plus or Free?

Not at launch. OpenAI plans to expand Dots to more users but hasn’t given a date.

Are Dots the same as ChatGPT Agent mode?

No. The old Agent mode has been discontinued. A Dot is a persistent assistant that can initiate work tasks across ChatGPT, Slack, Teams, and voice.

The shift from “chat on demand” to “delegated responsibility” is the real story here. Dots won’t replace how you use AI for quick questions — but for the ongoing, multi-step work that eats your week, having a coworker that stays logged in might be the most practical AI upgrade yet.