The GPT-6 rollout in ChatGPT
As of 2026-10-14, OpenAI has been rolling GPT-6 out in ChatGPT since 2026-10-07: the launch post says it began the day of publication in the Chat tab for the Plus, Pro, Business and Enterprise tiers, and that starting the next day (2026-10-08) the rollout expands to the Free and Go tiers, with Enterprise visibility depending on workplace admin settings. In the Chat itself, the four paid tiers run on GPT-6 Sol and Free and Go run on GPT-6 Luna, and OpenAI states that the models powering Work and Codex are not changing. The same post introduces Intelligent UI.
Updated 2026-10-14
One key across Claude, GPT and Chinese models, billed per token.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard / AMEX), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Four things to take away
Paid tiers / Free and Go
The post was published 2026-10-07: rollout began that day in the Chat tab for Plus, Pro, Business and Enterprise, and the post says it expands to the Free and Go tiers starting the next day, 2026-10-08.
The model behind the Chat
OpenAI's own wording: GPT-6 in ChatGPT is powered by GPT-6 Sol for Plus, Pro, Business and Enterprise and by GPT-6 Luna for Free and Go, with both said to be tuned for everyday conversation.
Work and Codex unchanged
The same paragraph closes by stating that the models powering Work and Codex are not changing as part of this release, and that the update applies to the Chat experience.
Weekly users, per the vendor
The post says GPT-6 is built for the more than 1.2 billion people who use ChatGPT each week. That is OpenAI's own figure, not independent measurement.
What actually changed
The Chat experience is what changed: GPT-6 arrives there together with Intelligent UI, which OpenAI describes as letting ChatGPT answer quickly with fully interactive user interfaces. Per the post, a response can include graphics, tappable buttons, forms, charts and interactive experiences you use directly inside the conversation, and the format depends on what you ask. OpenAI says it trained GPT-6 to decide how text, visuals and interactive elements fit together based on the question, while keeping the experience fast and familiar on web and mobile. For questions that need web search, the company claims GPT-6 Instant starts answering 44% sooner on average than GPT-5.6 Instant — OpenAI's own figure; the post does not say how it was measured.
OpenAI's own wording
The Availability section of the 2026-10-07 post reads verbatim: “GPT-6 with Intelligent UI starts rolling out globally to ChatGPT Plus, Pro, Business and Enterprise tiers today in the Chat tab. Starting tomorrow, the rollout expands to Free and Go tiers. Enterprise availability depends on workplace admin settings.” The next paragraph names the models: GPT-6 Sol for Plus, Pro, Business and Enterprise, GPT-6 Luna for Free and Go, both tuned for everyday conversation, the update applying to the Chat experience while the models powering Work and Codex are not changing. In that quote “today” is the publication date 2026-10-07 and “tomorrow” is the following day.
How GPT-6 reached ChatGPT
The API changelog releases GPT-6 Sol (gpt-6-sol) and GPT-6 Luna (gpt-6-luna): both accept text and image inputs and generate text through the Responses and Chat Completions APIs.
The changelog adds GPT-6.1 Sol (gpt-6.1-sol) for complex coding and professional work, officially described as lower cost than GPT-6 Astra, which the entry of 2026-09-03 lists as OpenAI's most capable model.
OpenAI publishes “GPT-6 and Intelligent UI for everyone”: the rollout starts in Chat that day for Plus, Pro, Business and Enterprise, the post says it expands to Free and Go from the next day, and the changelog the same day points chat-latest at the latest model those tiers use in ChatGPT.
Confirmed vs not published
Confirmed (checkable verbatim)
Checked 2026-10-14 and verifiable word for word on OpenAI's own pages: the publication date and the Availability section (tiers, the Chat tab, next-day expansion to Free and Go, Enterprise depending on admin settings); Sol powering the four paid tiers and Luna powering Free and Go; the sentence that the models powering Work and Codex are not changing; the description of Intelligent UI; and on the API side gpt-6-sol with gpt-6-luna dated 2026-09-22, gpt-6.1-sol dated 2026-09-29 and the chat-latest update dated 2026-10-07. The 1.2 billion weekly users and the 44% figure are the vendor's own statements, quoted as such.
Unverified and not published
The post publishes no usage limits for the Free and Go tiers. It gives no completion date for the rollout, only where it starts. The changelog says chat-latest points at the latest model used in ChatGPT by Plus, Pro, Business and Enterprise users but never names the snapshot, so it cannot be equated with gpt-6-sol. Whether Intelligent UI later reaches Work or Codex is unpublished. The 44% figure is OpenAI's own statement; the post does not say how it was measured. Note also that the Codex model page in the developer docs covers which models you can select in Work and Codex, which is a different question from which model powers the Chat; the two pages are quoted separately below.
The four paid tiers vs Free and Go
Plus, Pro, Business, Enterprise
The Chat runs on GPT-6 Sol, which the post says is tuned for everyday conversation. The developer docs add that these plans, plus Edu, can select GPT-6 Sol by ID (gpt-6-sol) in Work and Codex. Whether Enterprise sees it depends on workplace admin settings.
Free and Go
The Chat runs on GPT-6 Luna, likewise described as tuned for everyday conversation. The developer docs say Free and Go plans can choose gpt-6-luna in the desktop app, and describe Luna as OpenAI's most efficient model for focused, high-volume tasks such as summarization, extraction and focused coding. Usage limits and the date the rollout finishes are not published, so no numbers are given.
How to check whether you have it yet
① In the Chat tab ask something that invites structure — a comparison, a plan, a process — because OpenAI says the format follows the question, so a plain text answer is still a normal Intelligent UI outcome. ② If you are on Business or Enterprise and see nothing, ask your workspace administrator: the post states that Enterprise visibility depends on admin settings. ③ Keep in mind the developer docs tie model visibility to rollout progress, sign-in method and client, so web, desktop and CLI can show different lists. ④ On the API side, do not treat chat-latest as a production model: check the official model pages and name an explicit ID.
On QCode
ChatGPT tiers and API model IDs are two different things. As of 2026-10-14 QCode has gpt-6.1-sol, gpt-6-sol and gpt-6-astra available. QCode does not offer gpt-6-luna; if you need that family, use gpt-6.1-sol, gpt-6-sol or gpt-5.6-terra. Two settings change: put https://api.qcode.cc/openai/v1 in the SDK's base_url and one of those IDs in the model field, leave the rest as it is. Billed per token; per-model rates are listed on /models.
Frequently asked questions
Which ChatGPT plans have GPT-6 now, and since when?
Since 2026-10-07. The launch post of that date says GPT-6 with Intelligent UI started rolling out globally in the Chat tab to the Plus, Pro, Business and Enterprise tiers, and that from the next day, 2026-10-08, the rollout expands to the Free and Go tiers; it also states that Enterprise availability depends on workplace admin settings.
What is the difference between the models on the paid tiers and on Free and Go?
The model behind the Chat: GPT-6 Sol for Plus, Pro, Business and Enterprise, GPT-6 Luna for Free and Go. That is the whole of what the post says about the split — it adds only that both are tuned for everyday conversation, and it publishes no usage limits for either group.
How do the GPT-6 names in ChatGPT map to API model IDs?
OpenAI publishes no mapping. The changelog entry of 2026-10-07 says only that chat-latest points at the latest model available in ChatGPT for Plus, Pro, Business and Enterprise users and recommends the GPT-6 family for production API use. The IDs released so far are gpt-6-sol and gpt-6-luna (2026-09-22), gpt-6.1-sol (2026-09-29) and gpt-6-astra (2026-09-03). In code, name an ID instead of guessing from a plan name.
Which GPT-6 can I call on QCode — is Luna one of them?
gpt-6.1-sol, gpt-6-sol and gpt-6-astra are available on QCode. QCode does not offer gpt-6-luna, so use gpt-6.1-sol, gpt-6-sol or gpt-5.6-terra for that family. Use the OpenAI-compatible endpoint https://api.qcode.cc/openai/v1, put the ID in the model field and keep everything else as it is. Billed per token; per-model rates are listed on /models.
How do I pay for and activate it from mainland China?
You self-serve in the console: add balance, pick a plan, use the key. The recharge page offers Alipay, WeChat Pay, credit or debit cards (Visa / Mastercard / AMEX), Apple Pay, Google Pay and cryptocurrency, covering many coins and networks — and the account is prepaid. Billed per token; per-model rates are listed on /models.
Do the models in Work and Codex change too, and what about Intelligent UI?
No. The Availability section says the update applies to the Chat experience in ChatGPT and that the models powering Work and Codex are not changing as part of this release; Intelligent UI is described there as the answer format of the Chat, and the post says nothing about it reaching Work or Codex. The developer docs separately list which models you can select in Work and Codex (GPT-6.1 Sol, GPT-6 Sol and GPT-6 Luna, with Plus and the other paid plans plus Edu directed to gpt-6-sol and Free and Go to gpt-6-luna in the desktop app) — the two pages answer different questions and should not be merged into one reading.
Sources
OpenAI's product post “GPT-6 and Intelligent UI for everyone” (published 2026-10-07), the API changelog and the Codex model page on developers.openai.com, the endpoint and payment pages on docs.qcode.cc, and the live qcode.cc pages linked here — all checked 2026-10-14. Only vendor-official text and QCode's own pages are used; no media reports, social posts or forum threads.
Want GPT-6 on the API side? Start with one key
Switch between gpt-6.1-sol, gpt-6-sol and gpt-6-astra on one endpoint, billed per token; after you top up you activate everything yourself, effective immediately.
Related reading
GPT-6.1 Sol API Guide: Pricing, 272K Rule and vs gpt-6-sol
What gpt-6.1-sol costs, the 272K whole-request rule, and where it differs from gpt-6-sol.
GPT-6 Sol released: gpt-6-sol specs and pricing ($2 / $10, 1,050,000 context)
The specs checkable on the 2026-09-22 release date: context length, maximum output and input and output prices.
Paying for an AI coding API with Alipay and WeChat: channel limits, fees and settlement 2026
Per-payment caps, fees and settlement times for Alipay, WeChat Pay, cards and crypto.
Checked on 2026-10-14 against OpenAI's official pages and the text on docs.qcode.cc; vendor pages change without notice, so the official pages govern. The model list and per-model prices on QCode follow /models. Figures such as 1.2 billion weekly users or 44% sooner are OpenAI's own statements, not third-party measurements.