OpenAI is pushing ChatGPT further into autonomous work with the launch of Dots, always-on AI agents designed to keep working toward a user’s goals even when the user is no longer actively chatting with them. The company unveiled Dots at its DevDay conference in San Francisco on September 29, alongside a new $500-a-month Pro 500 subscription and several other updates to its AI products.
Unlike a conventional chatbot that waits for a prompt, Dots are designed to take ongoing responsibility for tasks. OpenAI says each Dot has its own cloud computer and browser, can work with connected applications and can continue working in the background. The agents are powered by GPT-6 Astra and can learn from user feedback over time.
OpenAI’s official Dots announcement
What Can OpenAI’s Dots Do?
OpenAI describes Dots as agents that can handle multi-step work rather than simply generate responses. Users can give a Dot a goal, and the agent can determine the steps needed to make progress before returning to the user when a decision or review is required.
Dots can be accessed through ChatGPT and can connect with communication platforms including Slack and Microsoft Teams. OpenAI says they can also work across more than 4,000 apps through connected tools. The company has designed them to handle recurring work and follow up on tasks rather than requiring users to restart the conversation every time.
That changes the role of ChatGPT from a system that responds to instructions into one that can continue working after the initial request.
For example, instead of asking an AI assistant to prepare a project update and then returning later with another instruction, a user could give a Dot responsibility for tracking the project, gathering information and preparing updates. The user remains involved when judgment or approval is needed.
Dots are Not Limited to the $500 Plan
The launch of Dots coincides with OpenAI’s introduction of its new Pro 500 subscription, but the two products are not the same thing.
OpenAI is initially rolling out Dots to Pro users in eligible markets, as well as Business Premium users. Enterprise customers can access the feature through a beta program when their workspace administrator enables it. OpenAI says the consumer rollout excludes the European Economic Area, Switzerland and the UK at launch.
The company says access is being rolled out gradually, so eligible users may not see Dots immediately.
OpenAI Introduces a $500-A-Month Pro Tier
Alongside Dots, OpenAI introduced Pro 500, a new $500-per-month subscription positioned above its existing Pro offerings.
The new tier provides the highest usage limits among OpenAI’s personal subscriptions and includes access to Astra Ultrafast; a higher-speed version designed for demanding workloads. OpenAI also made changes to the existing $200 Pro plan, reducing certain usage allowances after a transition period.
The move gives OpenAI a much wider range of paid subscription levels, with the company targeting users who rely heavily on ChatGPT, Codex and other AI tools for professional or development work.
The $500 price also signals how OpenAI is positioning increasingly powerful AI systems: not simply as software subscriptions, but as tools capable of performing substantial amounts of ongoing work.
GPT-6.1 Sol Arrives Alongside Dots
OpenAI also introduced GPT-6.1 Sol, a lower-cost model aimed at complex coding, computer use and professional workloads.
According to OpenAI, GPT-6.1 Sol delivers performance close to GPT-6 Astra at a lower cost. Its API pricing is $2 per million input tokens and $10 per million output tokens, with lower rates available for cached input.
The model is positioned as a more affordable option for developers building AI applications and agentic workflows.
The Bigger Shift is from Chatbots to Agents
Dots represent a broader change in how OpenAI is approaching AI.
Traditional chatbots largely operate on demand: a user asks a question, receives an answer and decides what to do next. Agent systems are designed to take on a goal, use tools, perform multiple steps and return when human input is required.
OpenAI’s Dots are part of that shift.
The launch also places OpenAI directly in competition with newer always-on agent products, including Meta’s Muse. Reuters reported that OpenAI is targeting growing demand for AI agents that can perform work across applications with less continuous human supervision.
For businesses, the implications could be significant. An AI system that can maintain context, access approved applications, operate a cloud computer and continue working between conversations has a different role from a conventional productivity assistant.
At the same time, giving an AI agent more autonomy creates a greater need for permissions, controls and human review. OpenAI’s enterprise controls allow administrators to manage access to Dots, connected applications, cloud computers, browsers and other capabilities. Custom rules can also guide when an agent should act or seek confirmation, although OpenAI says those rules cannot override built-in safeguards.
With Dots, OpenAI is betting that the next stage of AI adoption will involve systems that do more than answer questions. The company is building toward AI that can take responsibility for ongoing tasks, work across software and return to people when their judgment is needed.
