GPT-6 in ChatGPT: what it means for brand visibility

This text is part of the guide Brand visibility in AI: how it works and how to measure it.

What exactly does GPT-6 change in ChatGPT?

OpenAI has described a change in in a statement dated 7 October 2026. Późniejszy model GPT-6.1 Terra z 26 października, przeznaczony do trybu Work, Codex i aplikacji na API, opisujemy w tekście GPT-6.1 Terra: co to za model i co zmienia dla marek. Key facts:

• The first GPT-6 models were made available to users on paid plans in September, and in October the new version will be rolled out to everyone – which, according to OpenAI, amounts to over 1.2 billion people a week,
• The Plus, Pro, Business and Enterprise plans use the GPT-6 Sol model, whilst the Free and Go plans use the GPT-6 Luna model,
• Answers may include graphics, buttons, forms, charts and interactive elements, and the model itself selects the form of the answer,
• For questions requiring a web search, GPT-6 Instant begins to respond on average 44% faster than the previous version,
• The change applies to the chat tab; the layouts in Work and Codex modes remain unchanged.

What is Intelligent UI and what does it mean for brands?

Intelligent UI is a new way for GPT-6 to construct responses in ChatGPT: the model builds them using text, images and interface elements. A comparison might appear in side-by-side columns, a travel itinerary on a map, or a calculation in the form of a calculator that the user can use straight away in the conversation. When plain text is sufficient, the model sticks to text.

For a brand, this means that the answer increasingly takes the form of a comparison table. Our assessment: what counts in a comparison table are comparable facts – namely, price, specifications, location and who the service is aimed at. A brand whose website states these details explicitly provides the journalist with material for such a column. A brand described in general terms has nothing to contribute to the table. This is the same conclusion that emerged in the Wapiennik „Węże” case”: Citations increased when the standard, specifications and price were added to existing pages. We refer to this area as AEO.

Does GPT-6 use web searches more often?

OpenAI does not state that it does so more often. It says that GPT-6 makes better decisions about when to look something up online and is more confident in finding information that supports its answer. In an internal assessment of difficult questions, the new model was more likely to get to the heart of the matter than GPT-5.6. OpenAI has published the details of the tests in the model’s system card.

For the brand, this serves as a reminder that a search-based response relies on what the model finds and is able to read. The page must be indexed and accessible to the ChatGPT search engine bot (OAI-SearchBot; description in OpenAI documentation) and have the content in HTML, because in Vercel analysis None of the major AI bots executed JavaScript. The text describes how the model breaks a question down into sub-queries and selects sources How AI chatbots choose the brands they recommend. Why a Google ranking alone is not enough, as explained by Visibility in AI and SEO.

Why can both free and paying users see other brands?

This is because, since October, they have been using two different models. GPT-6 Sol supports paid plans, whilst GPT-6 Luna supports the Free and Go plans. According to OpenAI, both are optimised for everyday conversation, but as they are separate models, they may select different sources and brands for their responses.

This has practical implications for testing. If you’re checking brand visibility on a paid account, you’ll see a different model’s response to that seen by a customer using the free version of ChatGPT. On top of that, there’s memory, chat history and location. How to set up a test so as not to skew the results is described in How to check whether AI recommends your brand.

What does the change in model mean for visibility measurements?

The implementation of GPT-6 in ChatGPT marks a break in the data series. The September result and the October result come from different models, so the difference between them may not be due to anything the brand has done. Moreover, AI responses vary even without a change in model: in a study by SparkToro and Gumshoe.ai ChatGPT produced the same list of brands in fewer than 1 in 100 attempts. Read more in the article Why does AI give a different answer every time?.

When reading the results following such a change:

• Compare the period before 7–8 October and the period after it separately, rather than comparing individual days with one another,
• keep it the same frozen query set, because changing the set at the same time as changing the model makes any interpretation impossible,
• Don’t jump to conclusions based on the first few days, as the roll-out is still ongoing and not all users are receiving the new model at the same time,
• Look at the percentage of responses mentioning the brand, rather than its position on the list.

Taking measurements every day allows you to see from which day the results began to shift. We explain in the text how often to take measurements in order to distinguish a change from noise How often should visibility be measured?, and why the effects of actions take time to materialise, in the text Why do changes take time?.

Does GPT-6 change anything about Gemini, Claude and Perplexity?

No. This is a change in ChatGPT. The other models have their own update cycles and their own sources, so the brand might see an improvement in ChatGPT in October, whilst remaining unchanged in Gemini. That is precisely why it is not worth judging visibility based on a single chat.

At Aelo, we measure visibility across four models: ChatGPT, Claude, Gemini and Perplexity. Measurements in the module Visibility They are compiled daily, and a full report is produced once a month. The results are fed into the knowledge graph Synapsis, and on this basis we draw up specific measures in Tickets. It’s up to you to decide which of these you’ll implement. Agencies that need to explain the fluctuations to their clients following the change in model will find more information on the website Aelo for agencies.

Try it for free

FAQ

Sources

• OpenAI, GPT-6 and Intelligent UI for everyone (7 October 2026): https://openai.com/index/gpt-6-for-everyone/
• OpenAI, GPT-6 system specifications (October 2026): https://deploymentsafety.openai.com/gpt-6-october
• OpenAI, robot documentation: https://developers.openai.com/api/docs/bots
• SparkToro and Gumshoe.ai, a study on the consistency of brand recommendations in AI (January 2026): https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/
• Vercel, analysis of AI robot behaviour (December 2024): https://vercel.com/blog/the-rise-of-the-ai-crawler