OpenAI announces the launch of GPT-5: the new model will combine key features of previous versions

OpenAI has officially confirmed its intention to launch GPT-5 in the summer. The new model will combine the key advantages of previous developments — the multimodality of GPT-4 and the logical capabilities of the “o” series. The company’s goal is to create a powerful, universal model that eliminates the need for users to switch between different versions, informsBleepingComputer.

“We want not just to create a new flagship model, but also to combine our series into one,” – said Romain Yue, head of development work at OpenAI.

The new GPT-5 model will become not just a language system, but a full-fledged intelligent interface with wide capabilities. It will be able to work with text, images, audio and video within one session — providing true multimodality. One of the key improvements will be long-term memory: GPT-5 will store important data from previous conversations, opening a new level of personalization.

Another important step will be to unify all OpenAI tools in a single environment — from code and image generation to web search and voice requests. Such integration will make interaction with the system more convenient for both developers and ordinary users.

Accuracy is also improved: GPT-5, according to company officials, has significantly reduced the number of factual errors thanks to an updated architecture and expanded database. It demonstrates improved logical thinking and is able to generate complex solutions on its own, bringing it closer to the level of autonomous agents like AutoGPT — able to function with almost no human intervention.