Close Menu
Versa AI hub
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

What's Hot

Ryght’s journey to empower healthcare and life sciences with expert support from a hugging face

June 9, 2025

Benchmarking large-scale language models for healthcare

June 8, 2025

Oracle plans to trade $400 billion Nvidia chips for AI facilities in Texas

June 8, 2025
Facebook X (Twitter) Instagram
Versa AI hubVersa AI hub
Monday, June 9
Facebook X (Twitter) Instagram
Login
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools
Versa AI hub
Home»Tools»Chatgpt gains agent function for complex research
Tools

Chatgpt gains agent function for complex research

By February 3, 2025No Comments6 Mins Read
Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
Share
Facebook Twitter LinkedIn Pinterest Email

Openai has released a powerful agent function that Chatgpt is online, complex and complex and multi -step research tasks. It is reported that this function, called deep -search, can take hours or days to human researchers in a few minutes.

Openai explains deep research as an important milestone for traveling to artificial general information (AGI).

“The ability to integrate knowledge is a prerequisite for creating new knowledge,” says Openai. “For this reason, Deep Research has taken an important step toward a wider goal than developing AGI.”

With Agent AI, Chatgpt can support complex research

Douplied Search can autonomously find, analyze, and integrate information from hundreds of online sources. According to Openai, this tool can provide comprehensive reports comparable to the output of a survey, as prompts from users are prompted only by users.

Drawing a function from the Vararinant of the “O3” model, which is scheduled to be released on Openai soon, is to release users from time -consuming labor -intensive information gathering. Deep Reality promises accurate and reliable results, even for the competitive analysis of streaming platforms, information -based policy reviews, and even personalized recommendations for new commuting bicycles.

Importantly, all outputs include complete quotes and transparent documents so that users can easily verify the survey results.

The tools seem to be particularly skilled in clarifying niches and intuitive insights, and have become valuable assets in the entire industry, such as finance, science, policy planning, and engineering. However, OPENAI assumes that average research is useful for average users, such as super -person’s recommendations and shoppers looking for specific products.

People post a lot of wonderful examples, but here is something fun:

I’m in a book today and I’m looking for an old NSX. I searched for the perfect one and spent many hours searching. I was trying to give up and was trying to do a deep research … I found it.

-Sam Altoman (@sama) February 3, 2025

This latest agent function works through Chatgpt’s user interface. Users simply select the “deep -search” option in the message composer and enter a query. Support files or spreadsheets can also be uploaded for additional context.

When you start, AI embarks a strict multi -step process. This can take 5-30 minutes to complete. The sidebar provides the latest information on the taken action and consults with the source. The user can continue other tasks and will be notified when the final report is ready.

The result is displayed in the chat as a well -documented report in the details. In the coming weeks, Openai plans to further enhance these output by embedding images, data visualization, graphs, and providing a clear context.

Unlike GPT-4O, which is in real time in multimodal conversation, deep research prioritizes depth and details. The ability to strictly quote the source and provide a comprehensive analysis makes it stand out. Focus on the insights of well -documented research grades from high -speed answers.

Built for real world issues

DEEP RSEARCH uses sophisticated training methods based on browsing and inference tasks in the real world of various domains. The model was trained through reinforcement learning, and a multi -step research process was automatically planned and executed, such as adaptively improving the approach as backtracks and new information became available.

This tool browses the files that the user is apploaded, uses Python to generate and repeat the graph, embed media such as generated images and web pages in answers, quotes accurate sentences or passage from the source. can. The result of this extensive training is a very talented agent to work on complex real world issues.

https://www.youtube.com/watch?v=jv-lpisnloo

Openai has evaluated deep research in a wide range of expert -level exams, known as the “last test of mankind.” This test is composed of more than 3,000 questions that covers topics from rocket science, linguistics to ecology and classics, and test AI’s ability to solve multifaceted problems.

The result was impressive, and the model achieved a record accuracy of 26.6 % in these domains.

GPT-4O: 3.3 % GROK-2: 3.8 % Claude 3.5 Sonet: 4.3 % Openai O1: 9.1 % Deepseek-R1: 9.4 % Deep study: 26.6 % (view + python tool)

The deep -search has also reached a new cutting -edge performance with the GAIA benchmark. This evaluated the AI ​​models for real questions that require inference, multi -modal style Ency, and tool use ability. Deep research has led to the leader board the top with a score of 72.57 %.

Restrictions and issues

Chatgpt’s deep -search agent AI function means a bold step, but Openai acknowledges that this technology is still in the early stages and there are restrictions.

According to Openai, this system is significantly reduced compared to existing GPT models, but also “hallucinations”, provide incorrect inference, and provide incorrect inference. I have it. In addition, we are facing the task of distinguishing between authoritative information sources and speculative content, and we are struggling to adjust the confidence level.

Reports, quotes minor format errors, and delays in tasks can also frustrate initial users. Openai is expected to improve these problems over time through more use and repetition improvement.

Openai is gradually developing its functions, starting with PRO users who can access up to 100 queries per month. Plus and team layer follow, and enterprise access arrives next.

Residents in the UK, Switzerland and European economics are still unable to access functions, but Openai is working on expanding these regions.

In the next few weeks, Openai will extend this feature to Chatgpt’s mobile and desktop platform. Long -term visions include the ability to connect to the subscription base or unique data source, further enhancing the output and personalization.

Looking ahead, Openai assumes that Openai will integrate deep research with the “operator”, an existing chatbot function that takes actual actions. With this integration, Chatgpt can seamlessly handle tasks that require both asynchronous online surveys and real world execution.

(Photo by John Schnoblich)

See: Suspicion of data theft by Microsoft and Openai Prove Deepseek

Do you want to know more about AI and big data from industry leaders? See AI & Big Data EXPO held in Amsterdam, California and London. Comprehensive events will be held in collaboration with other major events, including Intelligent Automation Conference, Blockx, Digital Transformation Week, Cyber ​​Security & Cloud EXPO.

See more about Enterprise Technology events and webiners equipped with this TechForge.

author avatar
See Full Bio
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
Previous ArticleKATTEB AI: Content generator editing and calibration for you
Next Article True innovation requires responsibility: 3 priority items in 2025

Related Posts

Tools

Ryght’s journey to empower healthcare and life sciences with expert support from a hugging face

June 9, 2025
Tools

Benchmarking large-scale language models for healthcare

June 8, 2025
Tools

Oracle plans to trade $400 billion Nvidia chips for AI facilities in Texas

June 8, 2025
Add A Comment
Leave A Reply Cancel Reply

Top Posts

Deepseek’s latest AI model is a “big step back” for free speech

May 31, 20255 Views

Doudna Supercomputer to Strengthen AI and Genomics Research

May 30, 20255 Views

From California to Kentucky: Tracking the rise of state AI laws in 2025 | White & Case LLP

May 29, 20255 Views
Stay In Touch
  • YouTube
  • TikTok
  • Twitter
  • Instagram
  • Threads
Latest Reviews

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

Most Popular

Deepseek’s latest AI model is a “big step back” for free speech

May 31, 20255 Views

Doudna Supercomputer to Strengthen AI and Genomics Research

May 30, 20255 Views

From California to Kentucky: Tracking the rise of state AI laws in 2025 | White & Case LLP

May 29, 20255 Views
Don't Miss

Ryght’s journey to empower healthcare and life sciences with expert support from a hugging face

June 9, 2025

Benchmarking large-scale language models for healthcare

June 8, 2025

Oracle plans to trade $400 billion Nvidia chips for AI facilities in Texas

June 8, 2025
Service Area
X (Twitter) Instagram YouTube TikTok Threads RSS
  • About Us
  • Contact Us
  • Privacy Policy
  • Terms and Conditions
  • Disclaimer
© 2025 Versa AI Hub. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.

Sign In or Register

Welcome Back!

Login to your account below.

Lost password?