OpenAi is launching a new more powerful and long-awaited artificial intelligence model called GPT-5 in an attempt to stay ahead of growing rivals in the US and China.
The GPT-5, which will be presented during an event broadcast live on Thursday (7), is designed to be more capable of programming, creative writing and reasoning about complex consultations. At a press conference with journalists this week, OpenAi CEO Sam Altman called the model “a great update” compared to previous models of the company. “For the first time,” he said, “it really looks like you’re talking to an expert on any subject.”
OpenAI will make GPT-5 available to free users and paid subscribers from ChatgPT starting this Thursday, with clients from the education and business sector receiving access the following week. Paid customers will have higher use limits with the tool.

Take your business to the next level with the help of the country’s leading entrepreneurs!
Read more:
The company based in San Francisco started the BOOM of AI Generative almost three years ago with the launch of ChatgPT, originally fed by an anterior model called GPT-3.5. Since then, the company has launched a number of increasingly sophisticated systems, including multiple options that mimic the process of human reasoning and AI agents who perform tasks on behalf of the user with little or no intervention.
For much of this year, there was intense speculation around the GPT-5, fueled in part by Altman himself. In April, he said the GPT-5 would be “much better” than the company originally expected. More recently, Altman expressed some personal apprehension regarding the capabilities of the new model compared to his own. “I received an email with a question I didn’t understand very well and put it on the model-this is the GPT-5-and he answered perfectly,” Altman said in a podcast. “I felt useless about AI.”
Continues after advertising
The new model responds faster to users, is better in answering questions, and also invents less than previous models, said Nick Turley, who leads the chatgpt team at OpenAi. “When you talk to this thing, it looks a little more natural,” said Turley.
Read more:
As part of the release, OpenAI is introducing four predefined personalities that ChatgPT users can choose to better customize their interactions with chatbot. Options, initially available as “search preview”, include cynical, robot, listener and nerd.
Continues after advertising
With the GPT-5, ChatgPT can also determine when spending more time reflecting on user consultation and how long to dedicate to it, OpenAi said. Doing this automatically, instead of making users choose between a typical large language model or a reasoning system, can optimize the company’s computational resources and offer the most desirable response to users.
Crucially, OpenAI positions the model as especially useful for computer programming. A growing number of large technology and startup companies has launched AI assistants designed to help software developers at different levels to accelerate the writing and code clearance process, creating a “vibe coding” trend. “This on-demand software idea will be a definite part of the new age of GPT-5,” said Altman.
Read more:
Continues after advertising
In a live demonstration on Wednesday (6), a researcher showed how the GPT-5 could be used to quickly create a web app to help teach French to an English speaker. The researcher asked for a “beautiful and highly interactive” app with activities such as flashcards, quizzes and a classic game similar to Snake, where a rat eats cheese. Minutes later, ChatgPT generated the software, including the game.
To test the capabilities of the model, OpenAi has involved several companies as initial testing, including Anysphere, manufacturer of the popular AI coding assistant, and Lovable, a Swedish coding startup.
Anton Osika, co-founder and CEO of Lovable, said in an interview that his company found GPT-5 better than other AI models to quickly build complex applications, such as a financial planning app that includes a chatbot. Software developers typically need to itter during the construction of these applications, he said, and the new model makes it easier when producing easier code to maintain. It is also good for correcting bugs in software, he said.
Continues after advertising
Read more:
Although OpenAi has been a leader in the AI market, it faces renewed competition in some cases of former employees and allies. Google, from Alphabet, Anthropic and Xai, led by the Openai Con Musk co -founder, have launched models of reasoning and agents to rival OpenAi offers. They also focus on the rapidly expanding market for AI assisted coding. Chinese companies like Deepseek also gain space. And the goal invests heavily to hire talent for a new AI laboratory, including.
Hoping to maintain the advantage, OpenAi, valued at $ 300 billion, raised tens of billions of dollars to cover costs with talent, chips and data centers needed to build and support state-of-the-art models, including the GPT-5. OpenAi is also in initial negotiations on a possible sale of stocks to current and old employees, according to a report by Bloomberg News.
OpenAi’s not so secret weapon remains ChatgPT, chatbot that made the company known worldwide, which serves as an entry product for its various paid services. OpenAi said it now has almost 700 million chatgpt users per week and 5 million paying business users.
On Wednesday, as part of an effort to expand the adoption of Chatbot. OpenAi also launched two free and open models earlier this week to rival DeepSek and Goal offers.
© 2025 Bloomberg L.P.