OpenAI has launched deep research in ChatGPT, an AI-driven tool designed to conduct multi-step research on the internet. The new feature aims to help users complete complex research tasks in a fraction of the time it would take a person.
Deep research allows users to input a query, and ChatGPT will browse the web, analyze information, and generate a detailed report with citations. The tool is designed for professionals in finance, science, policy, and engineering, as well as individuals looking for thorough product comparisons.
“This feature is built for those who need reliable, in-depth research,” said OpenAI. “It finds and consolidates insights from across the web, ensuring users receive well-documented, verifiable information.”
Powered by OpenAI’s upcoming o3 model, deep research can analyze text, images, and PDFs while adjusting its approach based on the data it encounters. OpenAI describes this as a step toward its goal of developing AI that can produce novel scientific research.
Users can access deep research by selecting it in the ChatGPT message composer and entering their query. The tool provides a summary of its research steps and sources in a sidebar. Depending on the complexity of the request, it may take between 5 and 30 minutes to complete.
“Deep research is designed for cases where a simple answer isn’t enough,” OpenAI stated. “It gives users the ability to offload time-consuming research while ensuring transparency through citations.”
OpenAI plans to add features such as embedded images and data visualizations in the coming weeks. While GPT-4o remains ideal for real-time conversations, deep research is intended for tasks that require extensive information gathering and detailed analysis.