OpenAI has expanded access to its new AI research agent, Deep Research, which is now available to all paid ChatGPT users, including Plus, Team, Edu, and Enterprise plans. Initially launched for ChatGPT Pro users, Deep Research allows for more extensive querying, increasing from 100 to 120 queries per month for Pro users, while other plans get 10 queries each.
Deep Research is designed to conduct in-depth internet research, synthesizing information from numerous sources quickly, significantly reducing the time required for complex research tasks. Since its launch, enhancements have been made, including the capability to cite embedded images and improved support for analyzing user-uploaded files.
OpenAI positions Deep Research as a tool for professionals in fields like finance, science, policy, and engineering, as well as for consumers seeking personalized shopping recommendations. CEO Sam Altman praised it as one of his favorite features, emphasizing its efficiency in completing extensive research tasks in mere minutes.
In testing, Deep Research achieved a score of 26.6% on the “Humanity’s Last Exam,” a benchmark that signifies a substantial improvement in AI intelligence compared to previous models. However, OpenAI warns that users must verify the outputs, as AI-generated information can still include inaccuracies. This concern is highlighted by a legal case where a law firm was penalized for submitting fictitious information generated by an AI chatbot.