ChatGPT's new AI tool promises to save you hours of research

OpenAI, the company behind ChatGPT, recently introduced a new tool within ChatGPT, further advancing its efforts to make AI models more independent – essentially paving the way for AI agents. Now, it has launched yet another tool, continuing its push in the same direction.

OpenAI has rolled out a new tool in ChatGPT called Deep Research, designed to help users dive deeper into topics. This agent uses reasoning to sift through vast amounts of online information and handle multi-step research tasks, completing in minutes what would take a person hours.

In essence, Deep Research can create detailed reports by combining, analyzing, and synthesizing data from hundreds of online sources. Users can also use it to personalize shopping recommendations for major purchases that usually require a lot of research, like buying cars, appliances, furniture, and more.

– OpenAI, February 2025

Like most of OpenAI's newest features, Deep Research in ChatGPT is currently limited to paying Pro users. And even then, there's a cap – users can only run up to 100 queries per month.

Next in line for access are Plus and Team users, with Enterprise following after. Meanwhile, OpenAI is still working on rolling it out to users in the UK, Switzerland, and the European Economic Area.

That said, all paid users will soon see a boost in their usage limits. OpenAI plans to introduce a faster version of Deep Research powered by a smaller model that still delivers high-quality results. For now, Deep Research is only available on the ChatGPT web version, but mobile and desktop app support is expected within the month.

On paper, Deep Research looks promising and could be a game-changer for anyone who does heavy knowledge work. Based on what OpenAI has revealed, the tool seems to function a lot like Google Gemini's Deep Research feature.

Google introduced its version last year for paying users, offering a way to speed up research by delivering results in minutes instead of the hours it would typically take – pretty much the same goal OpenAI is aiming for.

That said, it's too early to tell which one is more accurate in real-world use, especially since OpenAI's Deep Research just launched and isn't widely available yet. Keep in mind, though, that while both tools can be useful, they're not flawless – hallucinations happen. So, I think it will be wise to not take everything at face value and always double-check your sources.