ChatGPT’s agent can now do deep research for you

Feb 03, 2025 07:11 AM - 6 days ago 8164

OpenAI has revealed different caller agentic characteristic for ChatGPT called heavy research, which it says tin run autonomously to “plan and execute a multi-step trajectory to find the information it needs, backtracking and reacting to real-time accusation wherever necessary.”

Instead of simply generating text, it shows a summary of its process successful a sidebar, pinch citations and a summary showing the process utilized for reference.

OpenAI heavy investigation sidebar connected ChatGPT

Image: OpenAI

Users tin inquire questions utilizing text, images, and further files for illustration PDFs aliases spreadsheets to adhd context, and past it will return “anywhere from 5 to 30 minutes” to create a consequence provided successful the chat window, pinch promises that successful the early it will besides beryllium capable to see embedded images and charts. OpenAI besides notes limitations for heavy research, saying it tin “sometimes hallucinate” and dress up facts, struggle pinch telling the quality betwixt charismatic info and rumors, and registry really definite it should complaint a response.

Developing ways for generative AI devices to beryllium much useful and worthy paying for is the early companies for illustration OpenAI have promised for agents, and it claims that heavy investigation is tin of operating astatine the level of a investigation analyst. The demo video included present originates pinch a petition for info connected changes successful the unit manufacture complete the past 3 years, pinch a consequence that includes slug points and tables.

This characteristic intimately follows OpenAI’s launch of Operator, a instrumentality that tin usage a web browser to complete tasks for you, and is akin to the Project Mariner investigation prototype Google showed disconnected successful December. Google’s instrumentality is not disposable to the nationalist yet, but heavy investigation is launching “with a type optimized for Pro users today.”

OpenAI is offering up to 100 queries per period for those paying the $200 monthly interest and “limited access” promised for Plus, Team, and eventually, Enterprise users, calling the expertise “very compute intensive,” requiring much conclusion compute the longer it takes to investigation something. It besides says that each paid users will get higher complaint limits successful the early erstwhile a faster, much cost-effective type is available.

A property merchandise says that the exemplary powering heavy investigation scored a caller precocious for accuracy connected an AI benchmark dubbed “Humanity’s Last Exam,” which asks for responses to expert-level questions. The OpenAI heavy investigation exemplary reached an accuracy of 26.6 percent pinch browsing and python devices enabled, good supra GPT-4o’s 3.3 percent, and the adjacent highest scorer, its o3-mini (high) exemplary evaluated only connected text, astatine 13 percent.

More