OpenAI has introduced anew AI peter called Deep Researchin ChatGPT , design to treat complex research tasks that would normally take hours if done manually . It is anAI Agentthat play much like Google Gemini ’s deep research cock . Here ’s a dislocation .
Instead of provide a straightaway solution like a distinctive chatbot , it searches the uploaded Commerce or the internet , analyzes , and summarise entropy from hundreds of source across the internet to make a elaborate report . So , if you want to memorize about a specific topic in particular , conduct market enquiry , or make big leverage decisions , you could use this characteristic and ChatGPT will generate a report . fundamentally , it summarizes everything on the internet related to that topic in a social organization that you desire .
For example , permit ’s say you require to compare the latest electric cable car . You could ask , “ rule the best electric car under $ 50,000 free-base on image , reliability , and charging speed . ” ChatGPT ’s Deep Research will crop multiple sources , filter out unreliable claims , and show a comprehensive breakdown of everything .
Interestingly , Google has a interchangeable AI tool with the same name , Deep Researchthat came out just two months ago . It ’s already useable on Google Gemini and work really well .
How to Use ChatGPT Deep Research AI Agent
Using Deep Research is uncomplicated . You type your query in ChatGPT and take theDeep Researchoption . If want , you may bind files or spreadsheets to render extra context .
ChatGPT will then ask a few doubt to well empathise your demand . Since the process can take anywhere from 5 to 30 minutes , this step help secure the report is better aligned with what you need .
Once you reassert the details , the tool starts mold in the backcloth — gather information , analyze sources , and resume key findings . You do n’t take to hold back though . ChatGPT will send a notification once the enquiry is complete . Useful as inscrutable research take time that you’re able to use to do some other work . The last output is a well - documented report that includes tables , bullet distributor point , and citation , make it easy to assert the information .
While the current version does n’t include images or other visual medium , OpenAI plans to add engraft images , graphs , and data visualisation in succeeding update to make the reports even clearer .
For now , Deep Research is available only toChatGPT Prousers ( $ 200 / month ) , with a limit of 100 inquiry per calendar month . PlusandTeamusers will get access in about a calendar month ( according to OpenAI ) , followed byEnterpriseusers . However , the shaft is presently unavailable in the UK , Switzerland , and the European Economic Area . OpenAI also contrive to release a more cost - effective version with increased query boundary for all paid users soon .
Also learn :
How ChatGPT’s Deep Research Works
Deep Research is an AI agent alike toOperator . It checks source , understands them , filters out unreliable information , and summarizes everything – without human intervention . The feature run on a different interlingual rendition of OpenAI ’s upcomingo3 model , optimized for web browsing and data analysis .
The model use reward learning to better over time and can even influence with user - uploaded file , patch graph , and press out perceptiveness from PDFs . In late AI performance tests , this manakin outperformed others , achieving 26.6 % truth on the challengingHumanity ’s Last Exambenchmark — far ahead of competitors likeGemini , Grok , and OpenAI ’s ownGPT-4o .
Despite its forward-looking capabilities , Deep Research is n’t perfect . The good example can still make mistake , misinterpret data , or struggle to separate authorised source from misinform information . It also make formatting errors in reports and citations . But OpenAI claim the simulation will amend over fourth dimension to address these military issue . In fact , Google ’s interpretation of Deep Research has the same challenges .