OpenAI has introduced a new AI agent capable of completing a multi-step research task independently in ChatGPT.
Called “Deep Research,” the AI agent runs on a version of OpenAI’s upcoming o3 model. It can “find, analyze, and synthesize hundreds of online sources to create a comprehensive report at the level of a research analyst,” OpenAI says.
The tool is built for people who do “intensive knowledge work in areas like finance, science, policy, and engineering and need thorough, precise, and reliable research.” At launch, it is available on the web for ChatGPT Pro subscribers ($200 per month), who can submit up to 100 queries a month.
To get started, eligible users can click the new Deep Research button in the message box and enter a prompt. In addition to text, users can add files or spreadsheets to provide the AI agent with better context.
Once it starts working, ChatGPT will open a sidebar summarizing the steps it takes and the sources used. Depending on the query, the tool might need five to 30 minutes to complete the task. After that, it will notify the user and provide them with a report in the chatbox.
“Every output is fully documented, with clear citations and a summary of its thinking, making it easy to reference and verify the information,” OpenAI said, adding that the tool “accomplishes in tens of minutes what would take a human many hours.”
Currently, the reports are text-only, but in the next few weeks, OpenAI will add embedded images, data visualizations, and other analytic outputs to improve clarity.
This Tweet is currently unavailable. It might be loading or has been removed.
On Humanity’s Last Exam, a recently released benchmark that evaluates the reasoning abilities of AI models, OpenAI’s Deep Research returned with 26.6% accuracy. In comparison, DeepSeek’s R1 model stands at 9.4%, and Google’s Gemini Thinking stands at 6.2%.
Interestingly, Google launched a similar tool with the same name last year for Gemini Advanced subscribers ($20 per month). One big advantage of Google’s model is that it lets users export the final report to Google Docs.
Recommended by Our Editors
OpenAI warns that its Deep Research tool might hallucinate. Users will also find “minor formatting errors in reports and citations” at launch, making it important to verify the provided information before taking any significant action.
While the AI agent is currently available to ChatGPT Pro users on the web, OpenAI expects to roll it out to mobile and desktop apps by the end of this month. As part of the staggered rollout, ChatGPT Plus and Team subscribers will receive it next, followed by Enterprise.
Deep Research’s announcement comes just over a week after OpenAI launched “Operator,” an AI agent that can automate web-based tasks like planning trips, ordering groceries, etc.
Get Our Best Stories!
This newsletter may contain advertising, deals, or affiliate links.
By clicking the button, you confirm you are 16+ and agree to our
Terms of Use and
Privacy Policy.
You may unsubscribe from the newsletters at any time.
About Jibin Joseph
Contributor
