OpenAI’s Latest Feature: Deep Research
What is Deep Research?
OpenAI has introduced a new agentic feature called Deep Research, which can operate autonomously to "plan and execute a multi-step trajectory to find the data it needs, backtracking and reacting to real-time information where necessary." This feature is designed to generate high-quality research results, similar to those produced by human researchers.
How Does Deep Research Work?
Instead of simply generating text, Deep Research shows a summary of its process in a sidebar, with citations and a summary showing the process used for reference. Users can ask questions using text, images, and additional files like PDFs or spreadsheets to add context. The model then takes anywhere from 5 to 30 minutes to develop a response, which is provided in the chat window.
Limitations of Deep Research
OpenAI notes that Deep Research can sometimes "hallucinate" and make up facts, struggle with telling the difference between authoritative information and rumors, and register how certain it should rate a response. These limitations are a step towards developing ways for generative AI tools to be more useful and worth paying for.
What Can You Expect from Deep Research?
Developing ways for generative AI tools to be more useful and worth paying for is the future companies like OpenAI have promised for agents. Deep Research is capable of operating at the level of a research analyst, as showcased in a demo video where a request for information on changes in the retail industry over the last three years received a response with bullet points and tables.
Comparison to Other Tools
Deep Research is similar to OpenAI’s launch of Operator, a tool that can use a web browser to complete tasks for you, and is also similar to the Project Mariner research prototype Google showed off in December. However, unlike Google’s tool, Deep Research is launching with a version optimized for Pro users today.
Pricing and Availability
OpenAI is offering up to 100 queries per month for those paying the $200 monthly fee and "limited access" promised for Plus, Team, and eventually, Enterprise users. The company notes that all paid users will get higher rate limits in the future when a faster, more cost-effective version is available.
Accuracy and Performance
A press release states that the model powering Deep Research scored a new high for accuracy on an AI benchmark dubbed "Humanity’s Last Exam," which asks for responses to expert-level questions. The OpenAI deep research model reached an accuracy of 26.6 percent with browsing and python tools enabled, well above GPT-4o’s 3.3 percent, and the next highest scorer, its o3-mini (high) model evaluated only on text, at 13 percent.
FAQs
Q: What is Deep Research?
A: Deep Research is a new agentic feature from OpenAI that can operate autonomously to plan and execute a multi-step trajectory to find the data it needs, backtracking and reacting to real-time information where necessary.
Q: How does Deep Research work?
A: Deep Research generates text based on user input, and provides a summary of its process in a sidebar, with citations and a summary showing the process used for reference.
Q: What are the limitations of Deep Research?
A: Deep Research can sometimes "hallucinate" and make up facts, struggle with telling the difference between authoritative information and rumors, and register how certain it should rate a response.
Q: Is Deep Research available to the public?
A: Yes, Deep Research is launching with a version optimized for Pro users today.
Q: How much does Deep Research cost?
A: OpenAI is offering up to 100 queries per month for those paying the $200 monthly fee and "limited access" promised for Plus, Team, and eventually, Enterprise users.
