google.com, pub-8701563775261122, DIRECT, f08c47fec0942fa0
USA

OpenAI launches new GPT-5 model for all ChatGPT users

In the picture, Openai CEO Sam Altman speaks with Softbank Group CEO Masayoshi Son at an event in Tokyo on February 3, 2025.

Tomohiro Ohsumi | Getty Images News | Getty Images

Openai announced on Thursday, the latest and most advanced large-scale artificial intelligence model GPT-5.

The company makes GPT-5 usable for everyone, including free users. Openai, especially in areas such as writing, coding and health services smarter, faster and more “much more useful”, he said.

“I tried to go back to GPT-4, and it’s quite miserable,” Openai CEO Sam Altman said in a briefing with journalists.

In 2022, Openai has become the mainstream since it launched AI Chatbot Chatgpt. The company said that as CNBC has already reported, it expects 700 million active users per week in Chatgpt this week and that it was negotiating with investors about a potential stock sale in a valuation of approximately $ 500 billion.

Openai said that the hallucination rate of the GPT-5 is lower, that is, the model produces the answers less frequently. The company said it made comprehensive security assessments while developing GPT-5, including 5,000 hours of test.

Openai will use the GPT-5 “secure completion” instead of refusing to answer users’ questions if the GPT-5 is potentially risky. This means that the model will give high -level responses within security restrictions that cannot be used to harm.

“The GPT-5, the GPT-5, was trained to explain the limitations more clearly when a task cannot be completed, that it cannot be completed, and this reduces the claims that are not supported compared to previous models,” said Michelle Pokrass. He said.

During the briefing, Openai showed how the GPT-5 could be used for “vibe coding”;

The company asked the GPT-5 to create a web application that can help a English speaker to learn French. The application had to have an interesting theme and include a way to monitor daily progress as well as activities such as flash cards and exams. Openai sent the same request to two GPT-5 windows and created two different applications within seconds.

Applications had “some coarse edges”, an Openai possible sales, but users can make additional settings to the software created by AI, such as changing the background or adding additional tabs, as they deem appropriate.

The GPT-5 is spreading to Openai’s free, pro, pro, pro and team users on Thursday. This launch will be able to access a reasoning model, which is a kind of model that makes a “thinking” or an internal chain of thought before responding to free users. If free users hit the usage cover, they can access the GPT-5 Mini.

Openai’s Plus users have higher usage limits and professional users have unlimited access to GPT-5 and access to GPT-5 Pro. Chatgpt EDU and Chatgpt Enterprise users will reach the GPT-5 for about one week from Thursday.

BoxA company that helps businesses manage computer files, has recently tested the GPT-5 in a wide range of data clusters.

Aaron Levie, the CEO of the Box, said that previous AI models failed in most of the company’s most advanced tests because they are struggling to make sense of complex mathematics or logic in long documents. However, Levie said GPT-5 is a “complete breakthrough”.

“The model can maintain more of the information it looks at and use a much higher level of reasoning and logic abilities to decide later,” Levie said in an interview. He said.

Openai releases three different versions of the model for developers through the application programming interface or API. These versions are designed for GPT-5, GPT-5-Mini and GPT-5-Nano, for different cost and delay needs.

At the beginning of this week, Openai released two open weight models for the first time since he launched the GPT-2 in 2019. These models were produced as lower cost options for developers, researchers and companies that can easily work and customize.

However, the GPT-5 and Openai’s wider consumer mass are also taken into consideration. The company said that interaction with the model felt natural and “more human”.

Altman said that the GPT-5 has a team of doctoral experts at any time.

“People are limited to ideas, but not really the ability to execute in many new ways,” he said.

Don’t miss this information from CNBC Pro

Related Articles

Leave a Reply

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

Back to top button