OpenAI dots: The Always-On Agent, Launch Details and API Status

What is OpenAI dots?

dots is OpenAI's always-on agent experience in ChatGPT. Instead of stopping after one answer, a dot can take ongoing responsibility for a goal, carry context between conversations and bring back work for review. It can use its own cloud computer and browser, write and test code, and work across apps you choose to connect.

OpenAI says its plugin ecosystem connects dots with more than 4,000 apps. You can interact through ChatGPT on desktop, web and mobile, with messaging integrations including Slack and Teams. Texting is a separate limited beta, not a capability available to every plan or region.

The earlier article covered a rumored agent under the name “o”. That name is historical, not the official launch name. The article now lives at /blog/openai-dots-agent; the old URL redirects here. dots is neither Meta Muse nor a renaming of the o1 or o3 reasoning-model family.

OpenAI dots release date and rollout

The official announcement date is September 29, 2026. OpenAI is gradually rolling out access, so the announcement does not mean every eligible account already has the feature. Its help page says rollout can take several days.

Plan	Launch availability	Qualification
ChatGPT Pro	Gradual rollout in eligible markets	Initially excludes the European Economic Area, Switzerland and the United Kingdom
Business Premium	Available across supported ChatGPT regions as rollout reaches accounts	Business Premium eligibility applies; do not assume every business plan includes it
Enterprise	Beta when enabled by the workspace administrator	Initially off by default; launch announcement also includes Edu and Healthcare

This eligibility summary was checked September 30. Check OpenAI's getting-started guide for changes rather than relying on an undated country list.

Create your first dot in the ChatGPT desktop app or desktop web at chatgpt.com/dots. After setup, you can message it in the mobile app when mobile access is available. OpenAI's help page says dots cannot currently be created on mobile and are not supported on mobile web.

What dots can do with a cloud computer

OpenAI's examples include turning customer feedback into tested code changes and pull requests, updating launch materials as a product changes, revising research analysis after new data, and preparing content for approval. These are vendor examples, not independent performance guarantees.

A dot can work on several projects and continue between check-ins. You can inspect its cloud computer, view in-progress or completed work, and schedule reminders or recurring checks. If you explicitly enable local-computer access, it can also work on that machine; local access is optional and starts off.

Do not interpret “always-on” as a guarantee that any task will complete correctly or run for an unlimited number of days. Start with a reversible task, review the result and use explicit boundaries before granting broader permissions. For example, ask for a draft research summary before authorizing actions that send messages or modify external systems.

GPT-6 Astra, Aeon rumors and the API boundary

OpenAI identifies GPT-6 Astra as the model powering dots. Both the launch announcement and help guide use that name. The earlier Aeon story is not the model name established by those sources, so this article no longer presents it as the product's confirmed foundation.

The model and the agent product are different layers. dots adds persistent context, a cloud computer, app permissions, background work and orchestration. Calling gpt-6-astra through a model API does not automatically reproduce that environment, and the dots announcement does not establish dots or aeon as a public API model ID.

For a custom agent, you still need your own task queue, storage, tools, retries, cancellation, spending limits and audit trail. Choose a documented API model for that workflow rather than substituting a product name into the model field.

dots pricing and usage: what is included?

Your first dot is included in eligible Pro and Business Premium plans at no extra cost. The launch announcement says conversations with your dot do not count against ChatGPT usage limits. Tasks it starts or manages in Codex or ChatGPT Work continue to count against those products' limits as usual.

Plans also include allowances for deeper work, with expanded launch-month allowances. The getting-started guide describes a first-month dots-usage promotion and says later plan terms will be shared. These statements should not be read as an unlimited-use promise: distinguish conversations, deeper dot work and delegated Codex or Work tasks, and check current plan terms before choosing a subscription.

OpenAI separately announced a new $500/month Pro tier for Ultrafast models and higher usage at DevDay. That is not a dots-only subscription or a requirement that every dots user buy the $500 tier; the first dot is included in eligible Pro and Business Premium plans. See OpenAI's GPT-6.1 Sol launch announcement. The dots announcement does not publish a dots per-token API price or announce a donut-shaped device as a shipped dots feature. Subscription inclusion is not the same thing as API billing.

Safety, connected apps and action approvals

OpenAI separates background proactive research from actions that change things. Proactive research uses read-only tools for connected apps; those tools cannot send messages, change app content or control a computer. Other actions are checked against instructions, custom rules and built-in safety requirements.

You choose which apps to connect, can review activity and can pause your dot. Local-computer access is off by default. Sensitive actions can require approval or a handoff; changing a password is an example OpenAI says must be handled by the user. dots can still make mistakes, including when following custom rules, so review consequential outputs.

For data handling, OpenAI says Business, Enterprise and Edu workspace content is not used to improve models by default; personal-plan users control the applicable training setting. Disconnecting an app does not erase information a dot already obtained. Read the getting-started guide and dots safety, security and privacy explanation before connecting sensitive systems.

How to use an OpenAI model through APIMaster today

For a programmable coding or multi-step workflow, create an APIMaster account, create an API key, set your OpenAI-compatible base URL to https://apimaster.ai/v1, and try gpt-6.1-sol:

curl "https://apimaster.ai/v1/chat/completions" \
  -H "Authorization: Bearer $APIMASTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6.1-sol","messages":[{"role":"user","content":"Inspect this failing test and propose a minimal patch."}]}'

This calls an API model, not dots. gpt-6-sol and gpt-6-astra remain alternatives after checking current model cards and route compatibility. Keep the model ID configurable and test tool support and safety controls before turning a one-shot request into an autonomous worker.

FAQ

Is OpenAI's agent called dots or o?

The official name is dots; an individual assistant is a dot. “o” was a name in pre-launch coverage, not the name in OpenAI's September 29 announcement. The old article URL permanently redirects to this dots page.

When did OpenAI release dots?

OpenAI announced dots on September 29, 2026. It is a gradual rollout for eligible Pro, Business Premium and Enterprise accounts, not immediate access for every user.

Who can use dots now?

Pro access initially excludes the EEA, Switzerland and the UK; Business Premium covers supported ChatGPT regions; Enterprise beta requires administrator enablement. Access can take several days to arrive. These are the September 30 help-page terms, not a permanent regional promise.

Is Aeon the model behind dots?

The official launch announcement and help guide identify GPT-6 Astra. They do not establish Aeon as the released product's model name or a public API model ID.

Can dots work in the background and use my computer?

Yes: a dot has its own cloud computer and can continue working between conversations. Your local computer is separate, and access is optional and off by default. Enable it only after reviewing permissions and the task's consequences.

Does dots require a $500 Pro Max subscription?

No. OpenAI did announce a $500/month Pro tier for Ultrafast models and higher usage, but not as a dots-only requirement. The first dot is included in eligible Pro and Business Premium plans without an extra charge; usage terms and delegated Codex or Work limits still apply.

Is there an OpenAI dots API on APIMaster?

APIMaster had 0 public dots routes in the September 30 check, and the reviewed launch material does not publish a dots API model ID. Use documented models such as gpt-6.1-sol, gpt-6-sol or gpt-6-astra for your own API workflow; they do not include the dots product environment.

Sources

• Introducing dots — OpenAI, September 29, 2026 — official launch name, GPT-6 Astra foundation, features and plan inclusion.
• Getting started with your dot — OpenAI Help Center — plan and region eligibility, desktop setup, computer access and controls.
• How we build safety, security and privacy into dots — OpenAI — security and data-handling boundaries.
• OpenAI API pricing — model API pricing, separate from dots plan access.
• APIMaster GPT-6.1 Sol card, GPT-6 Sol card and GPT-6 Astra card — active-route prices checked September 30, 2026.

For a custom API workflow at up to ~94.6% off the listed model prices in this check, register with APIMaster, create a key, and test gpt-6.1-sol through https://apimaster.ai/v1. To use dots itself, follow OpenAI's ChatGPT rollout and current eligibility documentation.