GPT-5 will have to wait – the new generation neural network will not be presented at the upcoming OpenAI DevDay

Last year, OpenAI held a major event in San Francisco where it announced a slew of new products, including the GPT Store as a sort of alternative to Apple’s App Store. This year the company has chosen a different presentation format – it will be a series of on-site events for developers. And you don’t have to wait for the announcement of the flagship artificial intelligence model – the company will talk about API updates and new services for developers.

Image source: Mariia Shalabaieva / unsplash.com

«We are not planning to announce our next model on DevDay. We will focus on educating developers about what is available and showcasing the stories of the developer community,” an OpenAI spokesperson told TechCrunch. This year’s OpenAI DevDay events will take place on October 1 in San Francisco, October 30 in London and November 21 in Singapore. As part of them, the company will hold seminars, sessions, demonstrations with OpenAI products and the company’s engineers, and presentations by the developers themselves are also expected. Registration costs $450 (free for some fellows), and applications close on August 15.

In recent months, OpenAI, which is expected to make major breakthroughs, has been limiting itself to small steps in the field of generative AI – the company is tuning and honing the work of its current flagship models GPT-4o and GPT-4o mini, reducing the frequency of failures. But for the same reason, it is beginning to lose its leadership in the industry, at least in certain indicators. One reason may be the lack of access to quality data for AI training – more and more resources are blocking access to web crawlers, and, according to some estimates, in the period from 2026 to 2032. AI developers may run out of data to train their systems.

To overcome this problem, OpenAI enters into agreements with publishers and data controllers to protect against potential copyright infringement claims. In addition, the company’s engineers are changing the architecture of AI systems – work is underway on a revolutionary project codenamed Strawberry: neural networks will learn to reason and solve mathematical problems. On the other hand, the change in strategy may indicate that OpenAI has become more serious about security issues – there is an opinion that previously the rapid release of new powerful products was more important to it.