ChatGPT Launches Deep Research Feature Amid DeepSeek Frenzy
OpenAI has introduced the Deep Research feature, enabling ChatGPT to synthesize and analyze information in just a few minutes for tasks that would otherwise take a person 30 days to complete.
According to OpenAI, Deep Research can autonomously plan, analyze, and execute multi-step processes to find relevant data, cross-check information, and update it in real time when necessary. Instead of merely generating text, this new ChatGPT feature displays the synthesis process, includes citations, and provides summaries of referenced sources.

ChatGPT compiles information and provides a real-time summary of the process with the new Deep Research feature.
Users can submit queries via text, images, or files such as PDFs and Excel spreadsheets. The system then takes “between 5 to 30 minutes” to generate a response. OpenAI has stated that, in the future, deep research content may be presented through images or more visually intuitive charts.
Kevin Weil, OpenAI’s Chief Product Officer, asserted that research tasks that previously required a person 30 days to complete can now be accomplished in less than 30 minutes using ChatGPT.
However, OpenAI acknowledges the limitations of Deep Research, warning that it may sometimes exhibit “hallucinations,” fabricate information, or struggle to distinguish between credible news sources and rumors.
In a demonstration video released by OpenAI, ChatGPT was tasked with summarizing changes in the retail industry over the past three years. The system quickly displayed the gathered content, sources, and synthesis process on the right side of the interface while presenting the final output in the main text editor.
The development team highlighted that Deep Research is beneficial for fields such as finance, science, engineering, policy-making, and even consumer research for purchasing decisions.
This new feature was launched just days after OpenAI introduced Operator—a tool capable of autonomously browsing the web to complete specific tasks for users, similar to Google’s Project Mariner, unveiled in December.
AI researchers classify new features like Deep Research and Operator as AI agents. OpenAI and similar companies see these AI agents as the future, aiming to make AI more practically useful and worth paying for, rather than just serving as an entertainment tool.

Deep Research is available in the ChatGPT Pro subscription plan for $200 per month.
OpenAI offers 100 Deep Research queries per month for users subscribed to the $200 Pro plan, with “limited access” for Plus, Team, and Enterprise users. Currently, the feature is only available on the web version, with mobile and desktop applications expected to follow later in the month. The company emphasized that this AI agent is “highly resource-intensive,” requiring significant computational power and longer processing times to conduct thorough research on a given topic.
The launch of Deep Research comes amid the rising popularity of DeepSeek, which has been making waves in the AI market over the past week with its diverse capabilities and lower costs for research and usage. The Chinese AI startup is considered a formidable competitor, pressuring OpenAI to continuously release new products to stay ahead. Over the past weekend, OpenAI also introduced o3-mini, an AI model designed for reasoning, described as “both powerful and cost-effective,” alongside new AI agents for ChatGPT.