Home » News » OpenAI Launches Deep Research AI Agent to Take On Google’s Gemini

OpenAI Launches Deep Research AI Agent to Take On Google’s Gemini

by Ravi Teja KNTS
0 comment

OpenAI has introduced a new AI tool called Deep Research in ChatGPT, designed to handle complex research tasks that would normally take hours if done manually. It is an AI Agent that acts much like Google Gemini’s deep research tool. Here’s a breakdown.

Instead of providing a quick answer like a typical chatbot, it searches the uploaded docs or the internet, analyzes, and summarizes information from hundreds of sources across the internet to create a detailed report. So, if you want to learn about a specific topic in detail, conduct market research, or make big purchase decisions, you can use this feature and ChatGPT will generate a report. Basically, it summarizes everything on the internet related to that topic in a structure that you want.

For example, let’s say you want to compare the latest electric cars. You could ask, “Find the best electric car under $50,000 based on range, reliability, and charging speed.” ChatGPT’s Deep Research will browse multiple sources, filter out unreliable claims, and show a comprehensive breakdown of everything.

Interestingly, Google has a similar AI tool with the same name, Deep Research that came out just two months ago. It’s already available on Google Gemini and works really well.

How to Use ChatGPT Deep Research AI Agent

Using Deep Research is simple. You type your query in ChatGPT and select the Deep Research option. If needed, you can attach files or spreadsheets to provide extra context.

ChatGPT will then ask a few questions to better understand your requirements. Since the process can take anywhere from 5 to 30 minutes, this step helps ensure the report is better aligned with what you need.

Once you confirm the details, the tool starts working in the background—gathering data, analyzing sources, and summarizing key findings. You don’t need to wait though. ChatGPT will send a notification once the research is complete. Useful as deep research takes time that you can use to do some other work. The final output is a well-documented report that includes tables, bullet points, and citations, making it easy to verify the information.

While the current version doesn’t include images or other visual media, OpenAI plans to add embedded images, graphs, and data visualizations in future updates to make the reports even clearer.

For now, Deep Research is available only to ChatGPT Pro users ($200/month), with a limit of 100 queries per month. Plus and Team users will get access in about a month (according to OpenAI), followed by Enterprise users. However, the tool is currently unavailable in the UK, Switzerland, and the European Economic Area. OpenAI also plans to release a more cost-effective version with increased query limits for all paid users soon.

Also Read:

How ChatGPT’s Deep Research Works

Deep Research is an AI agent similar to Operator. It checks sources, understands them, filters out unreliable information, and summarizes everything – without human intervention. The feature runs on a different version of OpenAI’s upcoming o3 model, optimized for web browsing and data analysis.

The model uses reinforcement learning to improve over time and can even work with user-uploaded files, plot graphs, and extract insights from PDFs. In recent AI performance tests, this model outperformed others, achieving 26.6% accuracy on the challenging Humanity’s Last Exam benchmark—far ahead of competitors like Gemini, Grok, and OpenAI’s own GPT-4o.

Despite its advanced capabilities, Deep Research isn’t perfect. The model can still make mistakes, misinterpret data, or struggle to separate authoritative sources from misleading information. It also makes formatting errors in reports and citations. But OpenAI claims the model will improve over time to address these issues. In fact, Google’s version of Deep Research has the same challenges.

You may also like