OpenAI Launches ChatGPT Agents
OpenAI has officially launched ChatGPT Agents, a new feature for ChatGPT that enables the AI to complete multi-step tasks autonomously, marking a significant evolution in consumer-facing artificial intelligence. The update is now available to ChatGPT Pro users, with gradual rollout planned for Plus and Team subscribers.
The new capability transforms ChatGPT from a conversational assistant into a full-fledged task-performing agent capable of operating a virtual browser, running code, filling out forms, interacting with websites, and generating files—all based on natural language instructions.
“This is the most capable agentic AI tool available to consumers today,” said an OpenAI spokesperson. “It coordinates reasoning, code execution, and real-world actions in a single flow.”
Key Features
ChatGPT Agents operate by coordinating multiple tools—including a browser, code interpreter, file system, and sandboxed terminal—to carry out complete tasks. Users describe the outcome they want, and the AI determines which tools to use, how to structure the steps, and when to request approval.
Tasks may include:
- Booking reservations after checking calendar availability
- Comparing products across multiple websites and creating summaries
- Generating and editing spreadsheets, presentations, and documents
- Reading and interpreting logs or datasets
- Filling out complex web forms
- Automating recurring processes
The system runs in a secure virtual environment and prompts users before executing irreversible actions such as submitting forms or downloading files.
Availability and Limits
| Tier | Status | Monthly Agent Tasks |
|---|---|---|
| Pro | Available now | ~400 interactions |
| Plus | Rolling out | ~40 interactions |
| Team | Rolling out | ~40 per seat |
| Free | Not supported | N/A |
Agents are currently unavailable in the European Economic Area and Switzerland, pending regional compliance updates.
The feature is accessible via the ChatGPT interface by selecting “Agent” mode or using the /agent command.
Safety and Oversight
OpenAI has implemented strict safety protocols:
- Agents must request user permission before taking major actions
- All steps are logged and visible to the user in real-time
- Sensitive actions like financial transactions are blocked
- Users can pause or cancel execution at any stage
According to OpenAI, the system was designed to prioritize transparency and user control, with all agent behavior restricted to the assigned virtual sandbox.
Industry Context
The launch comes amid a broader shift in the AI landscape from single-response chatbots to autonomous, multi-tool agents capable of acting across systems. Similar developments are underway across several AI companies, but OpenAI’s agent release is the most broadly deployed to date.
Early adopters are already using the new feature for business intelligence, content automation, research, scheduling, and customer support.
Outlook
OpenAI's ChatGPT Agent rollout positions the company at the forefront of consumer-facing AI autonomy. While the technology is still limited by safety protocols and usage caps, it lays the foundation for a future where AI systems can complete real-world tasks across apps and websites without direct user input for every step.
More details and usage guidelines are available on OpenAI’s Help Center.
Recommended
Inside Look at Using Claude Code Remote Control
Anthropic's Remote Control, currently in research preview, adds a third option for Claude Code users who need to step away mid-session. Your session keeps running on your machine while your phone becomes a window into it…
Read nextAbout the Author
Founder at EveryDev.ai
Apple, Disney, Adobe, Eventbrite, Zillow, Affirm. I've shipped frontend at all of them. Now I build and write about AI dev tools: what works, what's hype, and what's worth your time.
Comments
No comments yet
Be the first to share your thoughts