OpenAI Launches GPT-6 for All ChatGPT Plans

OpenAI began rolling out GPT-6 in ChatGPT's chat mode for Plus, Pro, Business, and Enterprise plans on Wednesday (October 7) and for Free and Go plans starting Thursday (October 8). Accompanying the model is the Intelligent UI feature, which allows the model to decide whether to respond in text or interactive elements: graphs, buttons, forms, and small on-the-fly tools, such as a savings calculator or a bill splitter. "GPT-6 and Intelligent UI, now rolling out in ChatGPT for everyone," the company stated on its official account on X.

Paid plans utilize GPT-6 Sol, while free plans operate with GPT-6 Luna. These two models are not new; OpenAI launched them on September 22 via the API, in Codex, and in ChatGPT Work. What changes now is the interface most users see. According to the company, the models used in Work and Codex stay the same with this update, which only impacts the chat experience.

The Model Chooses the Format

Until now, graphical interfaces within an assistant relied on prior programming: a plugin, an integrated app, a dashboard made by a developer. The Intelligent UI reverses this order. When asked about a travel itinerary, the model can plot the stops on a map; in a product comparison, create a side-by-side table; for a simple factual question, continue responding in a line of text.

For those selling software with a thin interface layer over data, this shift matters more than the model itself. A form generated by the assistant essentially serves as a data entry screen that today is someone else's product. The business logic and record system remain on the SaaS provider's side, but the interaction with the end user now occurs within ChatGPT, not in a separate app.

What Changes for the Corporate Administrator

In the Enterprise plan, availability depends on the environment administrator's configurations, according to OpenAI itself. This is where the announcement becomes relevant for the CISO. An interactive form within the conversation represents a new way to collect data inside a tool that, in many companies, was approved for text only. Leak prevention policies, audit trails, and retention rules were crafted with prompts and responses in mind, not with components that the model creates and the user fills out.

Distribution is staggered: paid plans receive the feature first, while Free and Go accounts gain access starting October 8. A day after the feature reaches corporate accounts, it will already be in the free version that employees use on personal devices, outside any company policy. In corporate accounts, OpenAI allowed the decision to release the feature to rest with the administrator, not the user.

Pricing and Scale

GPT-6 Sol costs $2 per million input tokens and $10 for output in the API, half of the $4 and $20 charges for GPT-5.6 Sol, according to the pricing table released in September. Luna is priced at $0.10 and $0.50. Putting these models in front of the free base of ChatGPT, and not just subscribers, is a bet that inference costs have dropped enough to support heavier interfaces at a consumer scale.

On the same day, Anthropic launched Claude Haiku 5.5, a small model priced at $0.10 per million tokens in prompts of up to 100,000 tokens, the same as Luna. The competition between the two companies is now happening on two fronts: token price aimed at developers, and response format targeted at end users.

An Interface Outside the Integrator's Control

For consultancies and integrators, the effect is significant. Internal assistant projects often involve weeks of interface development: dashboards, triage forms, visualizations for managers. If the general-purpose assistant starts generating these pieces on demand, the project value shifts to data integration, access controls, and governance, removing it from the visual layer.

There is a clear limit to what has been announced. OpenAI described consumer-driven examples such as revenue timelines, travel stops, and simple games, and did not present corporate cases or metrics for business use. The security of a model-generated component within a regulated flow, such as credit or healthcare, has yet to be documented by the company. It is within this gap that corporate buyers will test the feature in the coming months.