ChatGPT brings GPT-6 and Intelligent UI to free users
OpenAI rolled GPT-6 out to ChatGPT, with paid plans on October 7 and Free and Go on October 8. Free users get GPT-6 Luna and interactive answers called Intelligent UI.
3 min read

By the numbers
- weekly ChatGPT users, per OpenAI
- 1.2B+
- faster start on web-search answers, per OpenAI
- 44%
- when Free and Go users get GPT-6 Luna
- Oct 8
OpenAI is rolling GPT-6 out to everyone who uses ChatGPT. Paid plans got it on October 7, 2026, and Free and Go users get it on October 8, Resultsense reports. The release adds Intelligent UI, a feature that lets answers include charts, forms and buttons. For more than 1.2 billion weekly users, the default ChatGPT experience is changing.
Who gets which model
The two readable reports agree on the split:
| Plan | Model | Date |
|---|---|---|
| Plus, Pro, Business, Enterprise | GPT-6 Sol | October 7 |
| Free and Go | GPT-6 Luna | October 8 |
Android Headlines says Enterprise access depends on each workplace's administrators. GPT-6 replaces the "Latest" option in the model picker, and users can still switch back to GPT-5.6 Sol to compare answers. Neither report says how long the older model will stay.
Sol and Luna are separate models. Free users get the lighter one. Resultsense cites Anthropic's own benchmark table for Claude Haiku 5.5, which scores GPT-6 Luna at 16.4% on Terminal-Bench 4.0, against 39.2% for Haiku 5.5. That is a rival's chart, so treat it as one data point.
What Intelligent UI does
Intelligent UI lets the model decide when text is not the best answer. It can reply with interactive pieces such as charts, forms, buttons and small working tools, and it falls back to plain text when that is clearer, according to Resultsense.
The reports give these examples:
- a clickable diagram for a technical explanation
- a widget to scale a recipe for more guests
- a Sunday roast cooking timeline
- road-trip stops on a map
- a savings calculator, a group bill splitter and a small game
Resultsense says OpenAI built streamable interface components and a compiler that draws the interface piece by piece as the model writes. Android Headlines says the feature is optional and can be turned off in web settings. It is rolling out gradually on the web and in updated mobile apps, while older macOS and Windows desktop apps are left out for now.
OpenAI describes the aim as software that adapts to people "instead of people adapting to software," Resultsense quotes the company as saying.
Speed
GPT-6 can start answering while it is still thinking. Android Headlines says responses stream to the screen while the model keeps working in the background. For questions that need a web search, OpenAI says GPT-6 Instant starts replying 44% sooner on average than GPT-5.6 Instant. Both outlets present this as OpenAI's own figure, not an independent test.
Neither report gives usage limits for free accounts, or says which model modes Free users can pick.
What this means for developers
- Expect users to compare your app with ChatGPT. People who get calculators and forms inside chat may expect the same from your support bot or product assistant.
- Treat model names as moving targets. Sol, Luna and GPT-5.6 may all appear in a user's picker. If you test prompts against ChatGPT, record which model you used.
- Do not assume the free model matches the paid one. The Terminal-Bench figure is a rival's claim, but the gap between a small and a full model is real. Test on both tiers before you give advice to users.
- Verify OpenAI's speed claim yourself. A 44% faster start applies to web-search answers on one model pair.
The detail still missing is the free-tier usage limit. Until OpenAI states it, the real cost of the Luna model for heavy users stays unclear.
Sources
- OpenAI Is Finally Bringing GPT-6 to ChatGPT for Everyone - Android Headlines
- GPT-6 and Intelligent UI roll out to free ChatGPT users - Resultsense
Related articles

OpenAI Decisions API opens to all developers in beta
OpenAI's Decisions API is now in public beta. It returns a probability, a pick or a score instead of text, runs on GPT-6 Luna and costs $0.10 per 1M input tokens.

GPT-6.1 Sol nearly matches Astra at one-fifth the cost
OpenAI's new GPT-6.1 Sol cuts the factual-error rate to 7.7% and costs one-fifth as much as the shelved GPT-6 Astra flagship model.

GPT-6 Astra is generally available on Bedrock and Copilot
GPT-6 Astra is GA on Amazon Bedrock at $10 per million input tokens, and in GitHub Copilot for Pro+, Max, Business and Enterprise plans.
The weekly digest
One email every Friday with the week's top stories from all six desks: AI, dev tools, coding, the tech industry, startups and what's next. Free, no spam.