OpenAI rolls out GPT-6 with Intelligent UI in ChatGPT

The core advancement in this release is the Intelligent UI, which allows ChatGPT to move beyond text-only responses. By integrating native, streamable components and a compiler, the system can generate interactive elements like charts, buttons, and forms alongside prose. This aims to tailor the output format to the task at hand, potentially making the AI more practical for a wider array of everyday applications.

The new models also show speed improvements and enhanced reasoning. OpenAI's internal evaluations suggest that GPT-6 Extra High begins answering as quickly as GPT-5.6 Medium, while outperforming it overall on high-value agentic tasks. For web-search queries, GPT-6 Instant starts 44% sooner on average than GPT-5.6 Instant, and the company says the new generation more often identifies when to search and finds support for its answers.

Safety measures have also been updated, building on Astra advances to resist multi-turn attempts to bypass safeguards and address threats like cyberattacks and biological risks. Importantly, the models are trained to use conversational context for risk assessment and to provide clearer explanations when they lack necessary information or tools. This indicates an ongoing effort to balance capability with responsible deployment, even as the model's design judgment and experience range are noted as areas still needing work.

This rollout extends a model family first introduced to paying customers last month, signaling a broader strategic move towards AI that adapts its interface to user goals. OpenAI frames Intelligent UI as a step towards a future where users describe their desired outcome, and the AI constructs a bespoke experience around that task, aligning with its mission for general artificial intelligence.