OpenAI is bringing interactive visuals to ChatGPT with GPT-6

OpenAI has rolled out GPT-6 across ChatGPT subscription tiers, introducing an interactive interface that combines conversational responses with graphics, charts, and calculators. This update, announced on October 7, brings GPT-6 to Plus, Pro, Business, and Enterprise subscribers, while Free and Go users will begin receiving GPT-6 from October 8.

The availability depends on subscription arrangements, supported applications, and organizational settings. The Intelligent UI feature allows ChatGPT to determine the most suitable format for presenting information, such as diagrams, maps, interactive controls, or custom visual presentations. Users can compare products, explore scientific concepts, or engage with practical tools like personal finance calculators or bill-splitting apps within the same conversation.

The technology enables ChatGPT to create functional tools without requiring users to leave the conversation. OpenAI aims to move beyond text-based interactions to responses better suited for practical tasks. Educational applications, such as diagrams illustrating aircraft aerodynamics, and everyday planning tasks, like cooking instructions with adjustable ingredient quantities, will benefit from the new interface.

GPT-6 can also begin presenting information while continuing to reason or retrieve material, gradually adding findings in real-time. The system starts answering web-search requiring questions 44% sooner than GPT-5.6, on average. The interface relies on native components that can be displayed progressively as the model generates responses, allowing interactive elements to appear alongside written explanations without waiting for a separate application.

Enterprise administrators control model availability within their organizations. OpenAI has also strengthened safety measures against bypassing safeguards, deceptive behavior, and attempts to circumvent restrictions. These updates are part of OpenAI's internal evaluations and do not guarantee the models are immune to manipulation or misuse.

The new capabilities are available through supported web and updated mobile applications, while older desktop applications do not support Intelligent UI.