OpenAI has introduced a new GPT-6 model in ChatGPT for more than 1.2 billion people who use ChatGPT each week. The company introduced the first GPT-6 models for paid customers last month. The new model introduces Intelligent UI, which allows ChatGPT to combine text, visuals, and interactive elements to answer questions through interactive interfaces.
GPT-6: Intelligent UI
GPT-6 can compose responses using text, visuals, and interactive elements, choosing how they fit together based on the question. With Intelligent UI, responses can include:
- Graphics
- Tappable buttons
- Forms
- Charts
- Interactive experiences that users can use directly within a conversation
The format depends on the question. A comparison can be presented side by side, while an explanation can use an interactive diagram. When a simple text answer is more useful, ChatGPT can provide one.
For everyday questions, ChatGPT can make a plan easier to follow by showing how its different parts fit together. For example, the timing for a Sunday roast can appear alongside the recipe, while stops on a road trip can be displayed on a map with notes about possible detours. Similarly, visual and interactive explanations can let users explore how something works, change an input to see what happens, or work through a problem step by step.
Users can also ask ChatGPT to create a tool for the task at hand. Examples include:
- A calculator to explore how savings could grow
- A bill splitter for dinner with friends
- A game that can be played within the conversation
How Intelligent UI works
OpenAI gave GPT-6 the freedom to create the most useful interface for each question while keeping the experience fast and familiar across web and mobile. The system uses a library of native, streamable components and a compiler that processes the interface as the model generates it.
The library provides each response with a familiar design foundation while allowing GPT-6 to determine how the components fit together. Meanwhile, the compiler allows the interface to appear progressively as GPT-6 generates it, without waiting for the entire response to be completed.
OpenAI also expanded its training methods to help GPT-6 make decisions about content, layout, visuals, and interaction. The company evaluated the interfaces it creates for:
- Clarity
- Usefulness
- Completeness
GPT-6 learned to use the component library and make design decisions about how to organize information clearly, when to use interactivity, and when a simple text response is enough. OpenAI says there is still work to improve the model’s design judgment and expand what it can create. Intelligent UI allows ChatGPT to combine UI, data, and actions instead of fitting every question into the same conversational format.
Faster responses while reasoning
With GPT-6, ChatGPT can begin answering while it continues to reason. Previously, reasoning models waited until they finished thinking before presenting an answer. GPT-6 instead interleaves reasoning and answering, so users can start receiving information while the model continues working on the response.
The model was trained to account for waiting time and progressively build an answer from the knowledge and findings available at each stage. It can compose multiple partial responses, with each adding useful information without unnecessary filler, while keeping the final response cohesive and factual.
In internal evaluations:
- Agentic tasks: GPT-6 Extra High began answering in the same amount of time as GPT-5.6 Medium and achieved a better overall score than GPT-5.6 Extra High on high-value, everyday agentic tasks.
- Web search speed: GPT-6 Instant started answering 44% sooner on average than GPT-5.6 Instant for questions requiring web search.
- Web search quality: GPT-6 made better decisions about when to search and more reliably found information supporting its answers.
- Difficult problems: GPT-6 addressed the key aspect of the user’s question correctly more often than GPT-5.6 in an internal evaluation.
Safety improvements
GPT-6 improves how it recognizes real-world risks, responds honestly, and respects the limits that people set. The model builds on several of Astra’s safety advances to better resist attempts to bypass its safety training, follow safeguards, and communicate what it can and cannot do relative to GPT-5.6 Sol.
OpenAI updated GPT-6’s safety training based on lessons from real-world use to strengthen protections against high-risk misuse involving:
- Cyberattacks
- Biological threats
- Violence
In adversarial testing, GPT-6 showed stronger resistance to attempts to bypass its safety training, particularly attacks that adapt across multiple turns. The model is also trained to respond safely in higher-risk scenarios while avoiding unnecessary refusals of harmless requests.
GPT-6 uses conversation history and relevant context to recognize risks that may not be apparent from a single prompt. OpenAI also ran alignment evaluations throughout training, assessing honesty and responses to safety restrictions.
GPT-6 is better at recognizing and communicating the limits of its own capabilities, including when it lacks the information or tools needed to answer. More detail on the results and OpenAI’s review is available in the system card.
Availability
GPT-6 with Intelligent UI is rolling out globally through the Chat tab for:
- ChatGPT Plus
- ChatGPT Pro
- ChatGPT Business
- ChatGPT Enterprise
The rollout expands to ChatGPT Free and Go on October 8. Enterprise availability depends on workplace admin settings.
GPT-6 in ChatGPT is powered by:
- GPT-6 Sol: Plus, Pro, Business, and Enterprise
- GPT-6 Luna: Free and Go
Both models are tuned for everyday conversation. The update applies to the Chat experience in ChatGPT, while the models powering Work and Codex are not changing as part of this release.
What comes next
OpenAI says its mission is to deliver the benefits of artificial general intelligence safely to all of humanity. GPT-6 brings its latest advances in intelligence to more than 1.2 billion people who use ChatGPT each week.
Intelligent UI is a step toward a ChatGPT experience that can shape an interface around what a user is trying to do. Over time, OpenAI wants ChatGPT to create more interfaces for everyday tasks, allowing users to describe what they want to accomplish and work with an experience built around it.
For decades, people have had to learn how to use software by navigating fixed interfaces designed around particular tasks. OpenAI says it expects this to change, with software instead adapting its interface to the person and what they are trying to accomplish.