Jakarta, Indonesia Sentinel — OpenAI has announced the launch of a new feature called “deep research” for ChatGPT users. Described as an agent that uses reasoning to synthesize large amounts of online information and complete research tasks, the feature offers to simplify the users research process.
According to The Verge on Monday (February 3, 2025), the feature is designed to enable AI agents to conduct in-depth research and present findings in a more structured format, not just as plain text but also with step-by-step details of the research process.
In a blog post published on Sunday (February 2), OpenAI stated that this new capability is designed for “people who do intensive knowledge work in areas like finance, science, policy, and engineering and need thorough, precise, and reliable research.”
The company also noted that it could be beneficial for individuals making “purchases that typically require careful research, like cars, appliances, and furniture.”
Unlike the previous version of ChatGPT, which simply generated text-based responses, the deep research feature provides a transparent view of the research process in a side panel. This includes citations and summaries that can be used as references.
OpenAI revealed that the feature can autonomously handle multi-step research, meaning ChatGPT can plan, browse data, and adjust its findings based on new information received during the research process.
Users can also provide context through text, images, or additional files such as PDFs and spreadsheets. OpenAI claimed that the research process can take between 5 and 30 minutes to complete, depending on the complexity of the request.
Capabilities and Weaknesses
Based on recent testing, OpenAI’s deep research model achieved an accuracy rate of 26.6% with browsing and Python tools enabled. This figure is significantly higher than GPT-4o’s 3.3% and the next best-performing model, o3-mini, which scored 13% when evaluated using text alone.
However, OpenAI also cautioned that deep research is computationally intensive, meaning more complex queries may take longer to process. Additionally, the feature has some limitations, including the potential for “hallucinations” or fabricating facts, and difficulty distinguishing between credible information and rumors.
Read Also:
Seven Killed in Philadelphia Plane Crash Involving Mexican Nationals, Including Child
The deep research feature is available to ChatGPT Pro subscribers, who pay a monthly fee of $200 and receive up to 100 research queries per month.
OpenAI said deep research is being rolled out to Pro users today, with support for Plus and Team users coming next, followed by Enterprise. The company also promised to introduce higher rate limits in the future to enhance user experience.
(Raidi/Agung)