Chat maker We have developed the latest agent function called Deep Research. This can conduct multi -step research on the Internet for complicated tasks.
According to Openai, this new agent can take several hours in a few minutes.
The company states that while users provide prompts, Chatgpt will search, analyze, and generate data from online source to create research -level reports.
The agents built on Openai’s future model O3 are excellent in web browsing and data analysis using a huge amount of text, images, and interpreting and interpreting PDFs from the Internet. I am.
“The ability to integrate knowledge is a prerequisite for creating new knowledge. Deep Research is a more important step for AGI development. This is. I’ve been thinking for a long time that I can create new scientific research. ”
This new feature is designed to support finance, science, policy, and engineering experts that require detailed research.
It is also useful for shoppers who are looking for excessive personal recommendations related to purchases of cars, appliances, and furniture. In particular, the company stated that all outputs were completely documented by quotes and inference summaries, so that users can easily refer to information.
Deep Rice usually helps find a niche information and data that usually needs a wide range of viewing, alleviates the complex research of a single query in order to save time and improve decision -making. Masu.
It takes 5-30 minutes to provide insights, but the user will be notified when you are ready. Openai has also added to implement images, data visualization and analysis immediately.
“Today, we will start the next agent deep -search. This is like a superpower. On -demand experts! Using the Internet, complicated surveys and reasoning, and we can return reports. In the X post, Sam Altman, Openai, states:
Recently, Openai has introduced an operator survey preview. This is a function that allows tasks independently on the web. Using a built -in browser, operators can navigate the website, enter, click, scroll, and execute tasks based on user orders to talk to the content. The operator, which is still in the early stages, represents the first step of Altman -led company to create AI agents that can manage tasks independently.
