ChatGPT Dots Explained: Agents vs One-Off ChatGPT

ChatGPT Dots are always-on agents inside ChatGPT. They keep working on your goals between chats, on their own cloud computer, instead of waiting for your next prompt the way a one-off ChatGPT reply does. OpenAI introduced them on September 29, 2026.

That difference is the whole story. ChatGPT answers when you ask. A Dot keeps going when you walk away. To see how Dots fit next to other agents and assistants, see our map of AI tools.

What ChatGPT Dots are

OpenAI describes Dots as always-on agents built to handle ongoing work. They launched on GPT-6 Astra, and since a November 3, 2026 update OpenAI also uses GPT-6.1 Terra, so your Dot can draw on different models for different parts of its work (OpenAI). Dots keep learning from your feedback and can connect to the apps you already use through ChatGPT's plugin ecosystem (OpenAI cites more than 4,000 apps).

You can talk to your Dot in ChatGPT on desktop, web, and mobile. You can also reach it in Slack and Microsoft Teams. Voice calls are part of the product, and texting is on the roadmap.

Your first Dot is yours to name and shape. Over time OpenAI wants teams of Dots working together, plus specialist Dots that companies can set up with their own identity and access for clear jobs inside an organization.

Day-to-day Dot vs one-off ChatGPT

Think of ordinary ChatGPT as a sharp colleague you ping when you need an answer. You open a chat, ask, get a reply, maybe iterate, then close the tab. The work stops when the conversation stops.

A Dot is closer to a junior teammate who already has context and a computer:

• It has its own cloud computer and browser, so it can keep working without sitting on your laptop.
• It can run several projects at once and take new tasks without forcing you to juggle separate threads for every step.
• It learns preferences over time: how you write, what good looks like, which standards matter.
• It can work in the background (OpenAI calls some of this proactive research) using connected apps in a read-only mode until you widen permissions.
• It brings you finished drafts and decisions that need your judgment, instead of dumping every intermediate step into one long chat.

OpenAI's own examples are ordinary work, not sci-fi: a developer Dot that turns recurring customer feedback into pull requests with demo videos; a launch lead whose Dot revises assets when scope changes; a scientist whose Dot updates figures when new data arrives; a sales lead whose Dot keeps a proposal current through technical review; a creator whose Dot turns an interview transcript into clips, show notes, and social drafts.

The pattern is the same each time. You set direction. The Dot moves the work forward. You review and approve the parts that matter.

What stays under your control

Dots are not set-and-forget without guardrails. OpenAI builds them on ChatGPT's existing protections and adds agent-specific ones:

• You choose which apps the Dot can use.
• Custom Rules let you allow, require approval for, or block specific actions.
• Auto-review checks actions that could change accounts or share information.
• Sensitive steps (like changing a password) stay with you.
• You can open the Dot's computer, watch Activity View, pause, or redirect anytime.

Conversations with your Dot do not count toward ChatGPT usage limits. If the Dot starts or manages tasks in Codex or ChatGPT Work, those tasks still use those products' normal limits.

Who can use Dots today

OpenAI is rolling Dots out to ChatGPT Pro, Business Premium, and Enterprise plans in eligible markets. Pro access is for users 18 and older outside the European Economic Area, Switzerland, and the UK. Business Premium works in supported ChatGPT regions, and Enterprise (including Edu and Healthcare) users can try the beta when a workspace admin turns it on, per the OpenAI Help Center. Your first Dot is included on Pro or Business Premium at no extra charge, with an allowance for deeper work.

At launch, OpenAI also gave Dots extended limits for the first month. Dots launched on September 29, so that launch period is over. For the limits that apply now, check the OpenAI Help Center and the ChatGPT release notes.

That plan gate matters for the day-to-day vs one-off split: if you are on a lower ChatGPT tier, you still have ChatGPT for one-off help. You do not yet get a persistent Dot that keeps working overnight.

When a Dot helps (and when ChatGPT is enough)

Reach for a Dot when:

• The job spans hours or days, not one reply.
• You want progress while you are in meetings or offline.
• The work needs connected apps, a browser, or a private computer session.
• You care about continuity: same preferences, same project memory, fewer re-explanations.

Stick with one-off ChatGPT when:

• You need a quick definition, rewrite, or brainstorm.
• There is no ongoing project to own.
• You do not want an always-on agent with app access yet.
• Your plan does not include Dots.

Most people will mix both. Ask ChatGPT for the sharp one-shot. Hand the multi-day thread to a Dot.

How this ties to AI fluency (without the jargon tax)

AI fluency here just means knowing which tool shape fits the job. Chatbots answer. Agents act. Dots are OpenAI's consumer-facing always-on agent shape inside ChatGPT.

If you can name that difference in plain language, you already avoid the two common traps: treating every agent like a chatbot (and then wondering why it needs approvals and permissions), or treating every chatbot like an agent (and then wondering why nothing moved while you slept).

Where Dots meet AI engineering

You use a Dot, but someone had to build it. That someone decides what the agent can see, which actions need your approval, and what happens when a step fails. If you've ever wondered why an agent asks permission or pulls the wrong file, you're already thinking like the people who design these systems.

That work has a name: AI engineering. If you're curious what the role looks like day to day, our guide on what an AI engineer does covers the skills and how people get started.

Soft next step

If you just got a Dot (or you are still on one-off ChatGPT) and want structured practice turning AI into a working automation around your real job, explore AI Fluency. No coding required, with live sessions and 1:1 mentorship.

For more breakdowns like this one, browse the AI tools hub.

Sources (verified)

• OpenAI, Introducing dots (September 29, 2026, with the GPT-6.1 Terra update dated November 3, 2026)
• OpenAI Help Center, Getting started with your dot (plans, age and region rules)
• ChatGPT product page, Dots

Facts above are limited to what those pages state, last checked November 4, 2026. Plan availability and limits can change; re-check OpenAI before you buy or roll out.