ChatGPT adopts a dynamic interface with images, charts, forms and buttons

OpenAI launches Intelligent UI for ChatGPT

On October 7, 2026, OpenAI announced the availability of its “Intelligent UI” update for ChatGPT Plus, Pro, Business and Enterprise subscribers worldwide. The novelty allows the model’s response to include images, diagrams, forms and touch‑enabled buttons without the need to open a new window or application. One day later, on October 8, the company extended the functionality to users on the free and Go plans, completing a staged rollout that now covers the entire ChatGPT user base.

How the new dynamic interface works

Intelligent UI is tightly coupled with the GPT‑6 Sol and Luna models, which were trained to recognise when a response would benefit from a visual or interactive element and to format it within the text. OpenAI has built a library of reusable components—buttons, charts, forms—and a compiler that renders them progressively as the model generates its answer. In this way, the user does not have to wait for the model to finish writing all the text before seeing the interactive element.

Published demos show a diagram of a seven‑speed bicycle with buttons that highlight each part when pressed, and a Mahjong tutorial where the tiles appear as scrollable images. In addition, users can request “on‑the‑fly” tools such as a retirement‑savings calculator, a retro‑style game, or a bill‑splitting utility, which appear as interactive widgets embedded directly in the conversation.

Subscription availability and underlying models

Subscribers on the higher tiers (Plus, Pro, Business and Enterprise) receive the intermediate GPT‑6 Sol model, while users on the free and Go plans use the more efficient GPT‑6 Luna variant. Curiously, the Pro tier, which runs on the GPT‑6 Astra model, does not support the Intelligent UI feature, and the GPT‑6.1 Sol version remains limited to Work and Codex products, not being included in the general chat update. This segmentation suggests that OpenAI is testing the new UI with different model configurations before a universal rollout.

Perceived advantages and limitations

According to OpenAI’s communications, GPT‑6 can provide partial answers while continuing web searches or other tool‑based tasks, reducing perceived latency on complex queries. The company also claims that the new model shows greater resistance to attempts at bypassing its safety training compared with earlier versions. When a question can be satisfactorily answered with plain text alone, ChatGPT continues to deliver a text‑only response, avoiding unnecessary insertion of visual elements.

OpenAI has not published quantitative data on how visual responses affect token consumption, latency, or user satisfaction. It also has not specified UI performance on low‑cost devices or limited‑bandwidth connections, leaving the experience on less powerful setups uncertain. Moreover, the firm has not publicly addressed potential privacy and security risks associated with forms and buttons that could collect user data, nor detailed to what extent third‑party services or APIs can be integrated into the generated widgets.

Implications for users and developers

For users, the ability to receive interactive diagrams, built‑in calculators and mini‑games within the same conversation thread opens new possibilities for productivity and learning without switching applications. Developers building on the ChatGPT platform will be able to leverage the component library to create richer experiences, although they will need to consider the limitations of plans that do not support Intelligent UI. The emergence of forms and buttons also raises questions about data management and the need to validate information that the model solicits from the user.

In summary, the Intelligent UI update transforms how ChatGPT delivers information: it is no longer limited to static text but blends written answers with visual and interactive elements that appear in real time. Users can now expect diagrams with buttons, updatable charts and embedded utilities directly in the conversation, prompting a rethink of both how questions are phrased and how responses are interpreted. OpenAI has indicated that the functionality will continue to expand, but the long‑term effects on usage patterns and adoption remain unclear.