From blinking cursor to living tool chest
ChatGPT was born as a chatbot. For nearly a decade, the core experience remained the same: type into a text box, receive paragraphs of response. That's how humanity learned to converse with AI — and it worked. But the plain-text chat interface, which defined the entire category, is now reaching its end of life. With the launch of GPT-6, OpenAI is introducing "Intelligent UI," a paradigm shift that transforms every ChatGPT response into an interactive environment with graphics, buttons, editable charts, and tools the user can manipulate directly.
This isn't an aesthetic change. It's functional. Instead of describing how an airplane wing generates lift in a paragraph, ChatGPT now generates diagrams the user can explore in real time. Instead of listing recipe ingredients, it presents an editable form. Instead of showing a static number in a table, it creates an interactive bar chart the user can adjust. It's the same principle that turned Excel into a software category: when the presentation becomes interactive, information stops being passive.
How Intelligent UI works
According to Aarush Selvan, product manager at OpenAI, the goal is clear: "the most helpful answers aren't just text." The GPT-6 model analyzes the intent behind the question and automatically adapts how it presents the answer. If the user asks about comparisons, the model presents items side by side. If the question involves numerical data, it generates interactive charts. If it involves calculations, it creates a custom calculator for the task at hand.
The examples OpenAI demonstrated to journalists are concrete: a college student asks how an airplane wing generates lift and receives diagrams showing airflow in real time; someone planning a multi-day hike sees a map with interactive routes; anyone wanting to save money gets a personal savings calculator built right into the conversation. All of this happens without leaving the chat.
The feature is customizable. As with other aspects of ChatGPT's personality, users can reduce the amount of visual elements in responses if they prefer something more traditional. This flexibility matters: OpenAI understands that there's no single right way to interact with an AI, and Intelligent UI is an option, not an imposition.
The visual AI race: Google vs OpenAI
OpenAI isn't alone in this direction. Google launched a similar feature in May 2026 with Gemini, called "Neural Expressive," which also integrates visual and interactive responses into the conversation. Google described Neural Expressive as a new "functional design language" for the AI era, with fluid animations, haptic feedback, and vibrant colors.
The competition between these two approaches is interesting because it reveals two different philosophies. Google starts from visual design — the presentation is what defines the experience. OpenAI starts from functionality — the presentation is the means to make information actionable. Both converge on the same point: AI responses can no longer be just text.
This migration isn't accidental. The plain-text chat interface, which was the revolutionary discovery of generative AI, has already shown clear limitations. For simple questions, it works well. But for complex tasks — analyzing data, planning projects, solving technical problems — plain text is insufficient. Users end up asking the AI to generate code, open a spreadsheet, or create an external chart. Intelligent UI eliminates that friction by bringing those tools into the conversation.
Responding while "thinking": 44% less wait time
Beyond Intelligent UI, GPT-6 introduces a capability that radically changes the user experience: the model can begin responding to the user while it's still "thinking." In the traditional architecture, users wait for the model to complete all reasoning before seeing anything. With the new streaming-thinking mode, ChatGPT shows the first draft of the response while continuing to refine reasoning in the background.
OpenAI claims this approach reduces wait time by 44 percent, and in internal tests, GPT-6 outperformed its predecessor GPT-5.6 on complex web searches using this method. For the average user, the difference is noticeable: instead of three to five seconds of blank screen before the first word appears, the response starts flowing almost instantly and improves continuously.
Paying users get GPT-6 Sol, the most powerful model. Free and Go users receive GPT-6 Luna, which is faster and more efficient but with slightly reduced capabilities. The rollout begins today for Plus, Pro, Business, and Enterprise users, with free users receiving the feature the following Monday.
What this means for the category
Intelligent UI is more than a product update — it's a signal that the entire category of AI assistants is maturing. In the early years, generative AI proved it could converse. Now it's proving it can act. The transformation from passive text to interactive elements is a qualitative leap comparable to the passage from static documents to electronic spreadsheets: it doesn't just change the presentation, it changes the relationship between the user and the information.
Whether this evolution will create a new category of tools — or render traditional chatbots with their text-box-and-paragraphs interface as obsolete as MS-DOS was for graphical interfaces — remains to be seen. What we do know is that GPT-6 is here to change how millions of people interact with AI every day. The interface is no longer the channel — it's the tool.
Sources: The Verge, TechCrunch, The Decoder
✓ Independent sources cross-checked and verified before publishing
