Openai responds to Deepseek Hype with Chatgpt “Research” customer


The Openai Deep Research, a new agent working in New Megpt, launched independently, in-depth web research, data analysis, and reports-tasks that will take hours of people or even days.

The tool, now available for professional users, is designed for professionals in financing, science, politics, and engineering, as well as anyone looking for "comprehensive, accurate and reliable research", according to Openai's 2 February. advertisement.

This version comes at an increasingly competitive time in artificial intelligence research, with China -based China Dibsic The script program fluctuated on the artificial intelligence economy, claiming that it can match the Chatgpt intelligence without a heavy weight price.

CEO of Openai Sam Altman Gyid The tool is on Twitter, described "Like Great Power; Experts at demand!"

Altman shared the state of personal use as deep research helped him find a rare NSX in Japan after he almost surrendered to search manually.

Users can access deep research by identifying them in the Chatgpt messages composer, inserting query, attaching files or data schedules for an additional context.

Artificial intelligence then analyzes hundreds of sources over five to 30 minutes, exposing the sidebar to track the search process before submitting a report that was fully cited.

Despite its advanced capabilities, deep research is not without dodges.

Openai admits that the model "can sometimes cheer the facts or make incorrect conclusions" and fights with a distinction between reliable sources of wrong information.

Currently, professional users can reach 100 information per month, where Plus and Team users are expected to get after that, followed by institution agents.

Openai is said to make a faster and more effective release to increase the limits of querying paid users.

To support its case, Openai referred to the performance of Deep Research in the last humanity exam, which is knowledge of the expert level in the AI ​​test across more than 100 fields.

The DEP Research model is 26.6 %, as it exceeded the Deepsek-R-1's 9.4 % to the GPT-4O from Openai, which only managed from 3.3 %.

Openai's launch timing is not a coincidence - last December, Google Appear Its "deep search" feature of its Gemini AI model for the advanced Gemini users.

Artificial intelligence giants are now closed in a battle to determine the future of discovering independent knowledge that artificial intelligence drives-far from the era of Chatbot, which started everything.

Smart in general Newsletter

Weekly journey narrated GEN, AI Tawylidi model.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *