OpenAI launches GPT-6 with Intelligent UI for ChatGPT

OpenAI is rolling out GPT-6 with Intelligent UI to paid users today and free users tomorrow, enabling interactive visuals in ChatGPT responses. The update, unveiled Wednesday, brings multimodal, custom interfaces directly into the chatbot for over a billion users.

GPT-6 with Intelligent UI is rolling out globally to ChatGPT Plus, Pro, Business and Enterprise users in the Chat tab. Free and lower-cost Go tier users will receive the feature on Thursday. The new experience is powered by OpenAI's recent GPT-6 model, with the Sol version for higher-tier subscribers and the more efficient Luna model for Go and free users.

Features and Examples

Intelligent UI includes interactive elements such as tappable buttons, customized calculators, interactive charts, and editable graphs. The update generates visual, custom elements when helpful for answering user questions.

An example showed a user prompt asking to break down the design of a seven-speed bike. In response, ChatGPT created a diagram of the bike with buttons highlighting parts like Wheels or Drivetrain. Another tester generated a San Francisco apartment affordability calculator with movable sliders to adjust income and expenses, showing what percentage of take-home pay different units would cost. In an airplane seat explorer tool comparing cabin layouts, users could tap any seat on the map for more details, with exit rows highlighted in green.

GPT-6 enables ChatGPT to combine text responses with diagrams, charts, forms, and more. Users can ask the chatbot to generate tools like a retirement savings calculator, a bill splitter, or interfaces for retro games. Another example shows ChatGPT teaching Mahjong by generating visuals of tiles and placing them in scrollable categories.

Model Capabilities

GPT-6 shows stronger resistance to attempts to bypass its safety training. It was trained to decide when an interactive format is more useful than plain text. When selected, it may combine written responses with interactive elements. The format of the answer depends on the question asked.

The model improves response speed by allowing ChatGPT to begin answering while continuing to think. It performs better on web searches than GPT-5.6 Instant and starts answering search-based questions sooner on average. GPT-6 is also better at recognising and communicating the limits of its own capabilities.

Context and Implications

The visual results are similar to Google's 'generative UI' announced earlier this year for Search, which can output adjustable graphics. Chatbot experiences have increasingly become multimodal, blending text, images, and voice. OpenAI's launch signals that the multimodal era is here.

A larger shift toward generative user interfaces may spark an identity crisis for designers as AI dictates more of what software looks like for individual users. AI tools now able to quickly generate custom software make the future of chatbots more personalized. OpenAI states the goal is to make learning complex topics easier and suggests the feature may also make ChatGPT friendlier and more accessible.

Aarush Selvan, a product manager at OpenAI, said the design team spent time figuring out when adding diagrams, charts, or buttons adds value versus feeling cluttered. Selvan stated that ChatGPT has predominantly been a text-based interface and that the company wants to help users get things done, from everyday tasks to learning. He added, "If a user isn’t interested in visuals, they can tell ChatGPT to generate fewer of those outputs."

OpenAI staff demonstrated Intelligent UI in a call with journalists. The rollout will expand to Go and free tiers on Thursday. Free and Go tier users will receive the GPT-6 model and Intelligent UI update starting tomorrow.