The era of ChatGPT's endless walls of text is coming to an end. OpenAI has announced that its GPT-6 model family now includes an "Intelligent UI" capability.

With this update, the model won't just produce writing; it will generate visual responses featuring charts, clickable buttons, forms and dynamic user experiences.

Working interfaces instead of static replies

According to OpenAI, when the model detects that a response needs visualization or interaction, it will embed working interface components directly into the reply. In one demo, the parts of a 7-speed bicycle were visually broken down, with buttons added that reveal details when each part is clicked. Another example presented a roast lamb dinner plan: the menu, supported by food imagery, included an interactive button that automatically updated ingredient quantities when the number of guests changed.

ChatGPT will select the most suitable design itself from the broad component library it was trained on. Simple prompts still get text replies, but users can request a custom interface at any time.

Who gets it, and when?

2

The feature is available worldwide starting today to ChatGPT Plus, Pro, Business and Enterprise subscribers. Free and Go tier users will gain access starting tomorrow.

While the news has generated excitement, it also raises two questions: rich interfaces could burn through token quotas far faster, and integrating visual elements could add new dimensions to the hallucination risk. Only time will tell how justified these concerns turn out to be.