Synthetic intelligence agency OpenAI has launched Deep Analysis, a brand new AI-powered agent inside ChatGPT that independently conducts in-depth internet analysis, analyzes knowledge, and compiles reviews—finishing duties that will take people hours and even days.
The software, obtainable now for Professional customers, is designed for professionals in finance, science, coverage, and engineering, in addition to anybody in search of “thorough, exact, and dependable analysis,” in response to OpenAI’s Feb. 2 announcement.
The discharge comes at a time of rising competitors in AI analysis, with China-based DeepSeek flipping the script on AI economics, claiming it may well match ChatGPT’s intelligence with out the heavyweight price ticket.
OpenAI CEO Sam Altman hyped the software on Twitter, calling it “like a superpower; specialists on demand!”
Altman shared a private use case the place Deep Analysis helped him discover a uncommon NSX automobile in Japan after he had almost given up looking out manually.
Customers can entry Deep Analysis by choosing it in ChatGPT’s message composer, inputting a question, and attaching information or spreadsheets for extra context.
The AI then analyzes tons of of sources over 5 to half-hour, displaying a sidebar monitoring its analysis course of earlier than delivering a totally cited report.
Regardless of its superior capabilities, Deep Analysis isn’t with out its quirks.
OpenAI concedes that the mannequin “can typically hallucinate info or make incorrect inferences” and struggles with differentiating authoritative sources from misinformation.
At present, Professional customers can entry 100 queries per 30 days, with Plus and Group customers anticipated to get entry subsequent, adopted by Enterprise prospects.
OpenAI is reportedly engaged on a quicker, more cost effective model to extend question limits for paid customers.
To bolster its case, OpenAI pointed to Deep Analysis’s efficiency on Humanity’s Final Examination, an AI benchmark testing expert-level information throughout 100+ fields.
Deep Analysis’s o3 mannequin scored 26.6% accuracy, far surpassing DeepSeek-R-1’s 9.4% and even OpenAI’s GPT-4o, which managed solely 3.3%.
The timing of OpenAI’s launch is not any coincidence— final December, Google rolled out its personal “Deep Analysis” function for its Gemini AI mannequin for Gemini Superior customers.
The AI giants are actually locked in a battle to outline the way forward for AI-driven, autonomous information discovery—a far cry from the chatbot period that began all of it.
Typically Clever Publication
A weekly AI journey narrated by Gen, a generative AI mannequin.
