OpenAI announced a new ChatGPT agent called Deep Research. As the name suggests, it conducts multi-step online research for complex tasks.
Let’s say you’re a student who needs to submit a report on pure- and mixed-gas sorption for glassy polymers.
The official website shows Deep Research wrote a comprehensive report on that subject, complete with formulas, explanations, and sources.
OpenAI says this new AI model helped a user save four hours of research time.
Deep Research is easy to use, just like OpenAI’s other models:
- Open your OpenAi account for ChatGPT.
- Then, select “deep research” in the message composer and type your query. Also, you may attach files or spreadsheets to add more details to your questions.
- Wait for five to 30 minutes until Deep Research finalizes your result.
While it’s running, a sidebar appears to summarize the steps it took and the sources it used.
Interestingly, this feature is similar to DeepSeek’s ability to explain its reasoning for providing answers.
DeepSeek is the Chinese AI model that shook the global AI industry by matching the performance of Western counterparts at a lower cost.
Learn more about DeepSeek here.
Nevertheless, Deep Research has limitations. Like other AI models, it can hallucinate facts.
It struggles to distinguish between reputable information from rumors and verify how trustworthy a source is.
Moreover, it requires a massive amount of computing power, so it’s only available for ChatGPT Pro users.
At the time of writing, only Pro users will have 100 monthly queries for Deep Research.