OpenAI finished rolling out GPT-6 Intelligent UI to every ChatGPT tier on October 8, 2026, one day after paid plans got it. Free and Go users run GPT-6 Luna; Plus, Pro, Business and Enterprise run GPT-6 Sol. OpenAI says the model is “built for more than 1.2 billion people who use ChatGPT each week,” and that GPT-6 Instant starts answering 44% sooner than GPT-5.6 Instant on search questions.
Key takeaways
- Free and Go tiers got GPT-6 Intelligent UI on October 8, one day after paid plans.
- GPT-6 Luna lists at $0.10 input per million tokens — 20x below Sol’s $2.
- OpenAI aimed the release at 1.2 billion weekly ChatGPT users.
What is GPT-6 Intelligent UI?
GPT-6 Intelligent UI is a ChatGPT capability that lets the model return working interface elements instead of plain prose. OpenAI’s release notes describe buttons, forms, charts, side-by-side comparisons and editable diagrams, with the model choosing the format per question. It began rolling out October 7 and reached the free tier October 8.
The sharper version: ChatGPT can now build a small piece of software inside the answer. OpenAI’s own examples are a savings-growth calculator, a dinner bill splitter and a playable conversation game.
That is a different product category from a chatbot reply. A calculator that a user can type new inputs into is a tool, and tools are what a large slice of the open web currently sells against advertising.
OpenAI also says GPT-6 answers while it is still thinking, building the response progressively rather than holding the user on a loading screen. The company reports that GPT-6 Instant starts answering 44% sooner on average than GPT-5.6 Instant on web search questions.
It is not entirely new
Per OpenAI’s release notes, ChatGPT already had interactive charts as of June and interactive modules covering more than 70 math and science topics as of March. Intelligent UI generalizes those hand-built modules into something the model assembles on demand.
Which model does each plan get?
Paid tiers in the Chat tab run GPT-6 Sol. Free and Go run GPT-6 Luna, which OpenAI says is also tuned for everyday conversation. The Pro thinking level runs GPT-6 Astra, and Astra does not support Intelligent UI at all. ChatGPT Work and Codex models are unchanged.
The tier split is where the money is. API list prices published on OpenRouter put a 20x gap between the free-tier model and the paid one on input tokens, and the same 20x on output.
| Model | Who gets it | Input / 1M | Output / 1M | Intelligent UI |
|---|---|---|---|---|
| GPT-6 Luna | Free, Go | $0.10 | $0.50 | Yes |
| GPT-6 Sol | Plus, Pro, Business, Enterprise | $2.00 | $10.00 | Yes |
| GPT-6 Astra | Pro thinking level | $10.00 | $50.00 | No |
Sol and Luna are not new models. They shipped September 22 for ChatGPT Work, Codex and the API. What changed this week is that they reached the Chat tab, which is where the 1.2 billion weekly users actually are.
How much does GPT-6 Intelligent UI cost OpenAI?
OpenAI has not published a free-tier compute figure, so the honest answer is that nobody outside the company knows. But the Luna-versus-Sol price gap tells you what OpenAI was solving for: serving an interface-generating model to a billion non-paying users requires the cheapest model it has.
Interface generation is not free. A chart, a form and a calculator are more output tokens than a paragraph, and output is the expensive side of the meter — $0.50 per million on Luna against $0.10 for input.
Routing the free tier to a model priced 20x below Sol is how that math survives. It also explains the one-day staggered rollout: paid tiers first, free tier only after the cheaper path was live.
Here is the skeptical note. OpenAI gave no benchmark showing Luna matches Sol on the hard part — deciding when an interactive layout helps and when it is clutter. The company’s own framing is that both are “tuned for everyday conversation,” which is a claim about chat quality, not about interface judgment. Free users are getting the feature on a model nobody has shown can use it as well.
Who wins and who loses?
The winner is OpenAI’s advertising surface. A response containing buttons, forms and charts is a richer inventory slot than a wall of text, and the company has already begun selling against it. The losers are the publishers whose traffic is a single-purpose utility page.
Calculator and converter sites are the clearest exposure. Search Engine Journal’s Matt G. Southern noted that the bill-splitter and calculator examples “use generic inputs and public math,” which makes them trivial to reproduce — and leaves converter pages competing with a tool built inside the chat.
That is a real revenue line. Mortgage calculators, unit converters, loan amortization tables and BMI pages monetize high-intent traffic at display rates. None of that survives a model that renders the same widget for free.
- Wins: OpenAI’s ad inventory, which gets interactive slots — see our coverage of ChatGPT visual ads.
- Wins: Nvidia and the inference supply chain, since interface tokens are still tokens.
- Loses: Utility-page publishers — calculators, converters, splitters.
- Loses: Thin-wrapper startups selling a single widget on top of an API.
- Mixed: Google, which has pushed generative layouts in Search and is simultaneously cutting its Gemini free tier to one model on October 9.
The Google contrast is the most interesting line in the story. OpenAI is widening what the free tier can do on the same day Google narrows what its free tier gets. Those are opposite bets on whether free users are an acquisition cost or a loss to be contained.
What it means for API buyers
Luna at $0.10 input and $0.50 output is aggressive pricing for a 1.05M-context model. For comparison, Gemini 4 Argon lists at $2 a million, and Mistral Large 4 prices well above its rivals. The cheap tier is now the competitive battleground, not the frontier tier.
What are the limits on Intelligent UI?
The feature is narrower than the announcement implies. Per OpenAI’s release notes, Intelligent UI works at thinking levels from Instant through Extra High, but not at the Pro thinking level, and not in the older ChatGPT desktop apps for macOS and Windows. It is rolling out gradually on web and in updated mobile apps.
Users can disable the “Layout and visuals” setting on the web, though OpenAI says some visual elements may still appear.
Sourcing is the gap that matters for anyone checking numbers. OpenAI’s announcement does not say how citations will be attached to a chart, a comparison table or a generated calculator. Web search answers can carry citations behind a Sources button where available. A generated amortization table with no visible provenance is a different object from a cited paragraph.
What to watch next
Four signals follow directly from this rollout.
- October 9: Google’s Gemini free tier narrows to a single model, the clearest side-by-side test of the two free-tier strategies.
- Rollout completion: OpenAI calls the web and mobile rollout gradual and has not dated full availability. Watch for Pro-thinking-level support and the older desktop apps, both currently excluded.
- Citation policy: OpenAI has published nothing on how sources attach to generated charts and tools. Any update here is the signal publishers should track.
- Luna benchmarks: No public numbers compare Luna and Sol on layout quality. The first third-party evaluation is the one worth reading.
Frequently asked questions
Is GPT-6 Intelligent UI free?
Yes. Free and Go users began receiving it on October 8, 2026, one day after Plus, Pro, Business and Enterprise. Free and Go run GPT-6 Luna rather than Sol.
What is the difference between GPT-6 Sol and Luna?
Both are tuned for everyday conversation and both support Intelligent UI. The visible difference is price: Luna lists at $0.10 input and $0.50 output per million tokens on OpenRouter, against $2 and $10 for Sol.
How much faster is GPT-6?
OpenAI says GPT-6 Instant starts answering 44% sooner on average than GPT-5.6 Instant on web search questions. It also says GPT-6 Extra High begins answering in about the same time as GPT-5.6 Medium while scoring higher than GPT-5.6 Extra High.
Does GPT-6 Astra support Intelligent UI?
No. Astra powers the Pro thinking level, and OpenAI’s release notes state that Intelligent UI is not supported there. Astra lists at $10 input and $50 output per million tokens.
Can I turn Intelligent UI off?
On the web you can disable the “Layout and visuals” setting. OpenAI notes some visual elements may still appear.
Which companies are most exposed?
Publishers whose traffic comes from single-purpose utility pages — calculators, converters, bill splitters — face the most direct substitution, because the generated version is free and sits inside the answer.
Is GPT-6 available in the API?
Yes. Sol and Luna reached the API, ChatGPT Work and Codex on September 22, 2026. This week’s change was Chat-tab availability plus Intelligent UI.
The bottom line
GPT-6 Intelligent UI is the first time a frontier lab has shipped software generation to a billion-user free tier, and the 20x price gap between Luna and Sol is the mechanism that made it affordable. The capability story is real; the quality story is unproven, because OpenAI published no evidence that the cheap model exercises good layout judgment.
For investors, the trade is not OpenAI — it is the utility-page publishers now competing with a free widget, and the inference supply chain that gets paid either way.
Free guide · 12 pages
AI Tools Every Investor Should Use in 2026
10 tools for research, stock signals, charts and crypto, what each one costs, three ready-made stacks from $0 a month and 7 copy-paste prompts for filings and earnings calls. Enter your email and we will send it to you.
Plus one email a week: AI Money This Week, every Sunday. Unsubscribe anytime.
Sources
- OpenAI — ChatGPT release notes, “GPT-6 and Intelligent UI,” October 7, 2026
- Search Engine Journal — “ChatGPT Gets GPT-6 And Intelligent UI For Interactive Answers,” October 7, 2026
- Business Today — “OpenAI rolls out GPT-6 to all ChatGPT users with ‘Intelligent UI’,” October 8, 2026
- OpenRouter — GPT-6 Astra, Sol and Luna API pricing
Wealth Engine researches and drafts with AI tools and checks every figure against the sources above. How we report.
Leave a Reply