Introducing Deep Research

OpenAI has launched Deep Research in ChatGPT, an agentic feature that performs multi-step research on the internet for complex tasks. This feature completes work that would take a person many hours in just tens of minutes, finding, analyzing, and synthesizing hundreds of online sources to produce reports at the level of a research analyst.

It is based on an optimized version of the o3 model, leveraging strong reasoning capabilities to search and interpret vast amounts of text, images, and PDFs. A key characteristic is that it flexibly adjusts its search direction when it discovers new information during the information-gathering process, continuing its in-depth analysis.

Key features and capabilities are as follows:

• Autonomous research and synthesis: Independently explores the web and derives insights from diverse sources.
• Transparent evidence: Includes clear citations and a summary of its thought process in every output, making it easy to verify information.
• Real-time progress tracking: A sidebar allows users to check research steps and sources used in real time.

Usage limits per user are capped at 25 queries per month for Plus, Team, Enterprise, and Edu users, 250 for Pro users, and 5 for Free users. Once the quota is used up, the service automatically switches to a lightweight version based on o4-mini, allowing continued use.