OpenAI has unveiled “Deep Research,” a new feature within ChatGPT designed to enhance the artificial intelligence platform’s research capabilities.
Currently exclusive to ChatGPT Pro users, Deep Research aims to facilitate detailed and comprehensive exploration of complex topics, ranging from academic subjects to market trends and technical analyses.
Deep Research offers an intuitive interface to streamline the research process. Users initiate their inquiries in a panel on the right side of the screen, while the results are displayed on the left.
This side-by-side layout allows for efficient interaction with the AI while simultaneously reviewing the gathered information. Depending on the complexity of the topic, the research process can take between 5 to 30 minutes, enabling the AI to delve deeper, analyze extensive data, and provide more detailed responses compared to standard queries.
Deep Research is powered by a version of OpenAI’s upcoming o3 model, optimized for web browsing and data analysis. Upon receiving a user prompt, ChatGPT utilizes this model to search, analyze, and synthesize information from various online sources, including text, images, and PDFs.
The result is a comprehensive report that emulates the work of a research analyst, accomplishing in minutes what might take a human several hours.
Current Limitations and Future Enhancements
While Deep Research represents a significant advancement in AI-driven research tools, it is not without limitations. The tool may struggle to distinguish authoritative information from rumors and often fails to convey uncertainty accurately.
Additionally, there is a limit of 100 research sessions per month for Pro users. OpenAI acknowledges these challenges and plans to expand this limit and improve the feature’s speed and cost in the future.
As of its launch, Deep Research is available on the web version of ChatGPT and is expected to roll out to mobile and desktop applications later in February 2025. This development follows OpenAI’s January preview of another AI tool called Operator, which can perform a variety of tasks such as creating to-do lists or assisting with vacation planning.
