OpenAI has confirmed that GPT-5, the long-awaited new AI model, will be available this summer. According to CEO Sam Altman, the model will be a combination of all previous innovations: a single powerful system that can handle text, images, audio, and possibly even video.
The announcement came during OpenAI’s first official podcast episode, in which Altman (photo) casually mentioned that the launch is likely to take place sometime this summer. This comes after months of speculation and anticipation within the AI community.
No more model selector needed
Where ChatGPT users currently have to choose different models for specific tasks, such as GPT-4o for text, DALL·E for images, or the fast GPT-4o for multimodal interaction, GPT-5 brings everything together in a single, unified interface. This also eliminates the need for a so-called model selector.
Romain Huet, Head of Developer Experience at OpenAI, previously stated that with GPT-5, the company is combining the best elements of existing models. The goal is a fundamentally new AI system that is more powerful. It is also simpler and more consistent to use.
Full multimodal capabilities
According to sources close to the company, GPT-5 will have larger context windows. These are essential for long conversations and complex tasks. The model will also respond in a more personalized way as standard. It learns from user interactions to tailor responses to individual preferences.
In addition, GPT-5 will have full multimodal capabilities. This means that the model can seamlessly switch between text, images, speech, and possibly even video without the need for separate tools or extensions.
Fierce competition
The pressure on OpenAI is intense. While the company was still the clear market leader in 2023 and 2024, competition increased significantly in 2025. Tech giants such as Google, Anthropic, and Meta invested billions in AI development. Google’s Gemini 2.5 Pro and the open-source DeepSeek R1 are attracting a lot of attention due to their strong performance on specific tasks such as mathematical reasoning.
At the same time, OpenAI recently lost some of its top talent to other parties. Meta, among others, is said to have actively poached researchers. However, this does not seem to have affected the upcoming release of GPT-5.
An analysis of OpenAI’s releases shows that the pace is fast: GPT-4 came out in March 2023, GPT-4-Turbo followed in November, and GPT-4o appeared in May 2024. GPT-5 is expected to continue this trend, but in a much more competitive playing field.
According to OpenAI, GPT-5 is not just a new version, but a fusion of everything the company has developed in recent years. Whether this will actually enable it to regain its lead in the AI market will become clear in the coming months.