OpenAI unveils new deep research agent in ChatGPT, helping in complex web research tasks

OpenAI has introduced a new feature in ChatGPT called deep research, an agentic capability designed to conduct multi-step research on the internet for complex tasks. This feature can perform in tens of minutes what typically requires many hours for a human, by finding, analyzing, and synthesizing information from numerous online sources. It is powered by a version of the forthcoming OpenAI o3 model, optimized for web browsing and data analysis, and is capable of reasoning, searching, and interpreting large quantities of text, images, and PDFs.

Deep research is tailored for professionals in fields such as finance, science, policy, and engineering who require detailed and accurate research. Each output includes full documentation with citations and a summary of the analysis process for easy verification. The process takes between 5 to 30 minutes, depending on the complexity of the task. OpenAI plans to enhance these reports with embedded images, data visualizations, and other analytical outputs in the coming weeks.

Currently available to Pro users with up to 100 queries per month, deep research will be extended to Plus and Team users next, followed by Enterprise users. It is accessible today on ChatGPT web and will be available on mobile and desktop applications within the month.