Thursday, Oct 8 | --:--
Back to home

GPT-6 Rolls Into ChatGPT, Paid Plans First and Free on Luna From Oct. 8 — and the Answer Becomes an Interface

OpenAI began rolling GPT-6 into ChatGPT on Oct. 7: GPT-6 Sol for Plus, Pro, Business and Enterprise from Oct. 7, GPT-6 Luna for Free and Go from Oct. 8. The models reached the API on Sept. 22, so this is a distribution event, not a model launch. The new piece is 'Intelligent UI', which builds charts, forms, buttons and small in-chat tools into answers. OpenAI's speed and quality figures come from its own internal evals and have not been independently tested in ChatGPT.

Times of AI Desk 5 min read San Francisco, CA View as Markdown
Cover illustration for GPT-6 Rolls Into ChatGPT, Paid Plans First and Free on Luna From Oct. 8 — and the Answer Becomes an Interface

GPT-6 is not new. Rolling it into ChatGPT, free tier included, is. OpenAI's Sol and Luna models have been on the API since September 22; on October 7 OpenAI began rolling GPT-6 into ChatGPT, starting with paid plans that day and reaching Free and Go users from October 8. The free tier is where most people actually meet its models. OpenAI says more than 1.2 billion people use ChatGPT each week, according to The Indian Express's account of the announcement.

The product bet riding along is bigger than the model swap. Intelligent UI lets ChatGPT answer with charts, diagrams, tappable buttons, forms and small working tools instead of only paragraphs. That turns the chat window into a surface that renders little apps on demand. Within the same hour, Anthropic priced Claude Haiku 5.5 to match Luna's API list price for short prompts. The cheap tier is now the competitive front on both the consumer and developer sides.

Who gets what

Plan Model in ChatGPT From
Plus, Pro, Business, Enterprise GPT-6 Sol October 7
Free, Go GPT-6 Luna October 8

The schedule comes from OpenAI's announcement as posted to its developer community forum; The Verge reports the same split, with free and Go users following on Thursday. OpenAI's post on X called it "GPT-6 and Intelligent UI, now rolling out in ChatGPT for everyone." Per The Indian Express, OpenAI says the models behind ChatGPT Work and Codex are not changing in this release; the update is to chat.

The Verge describes OpenAI as "bringing the Sol and Luna versions to ChatGPT." OpenAI has not said, in the material we could read, whether the ChatGPT models are identical to the API snapshots.

What Intelligent UI does

ChatGPT can now compose an answer from text, visuals and interactive elements and still fall back to plain text when that is more useful. The Verge lists diagrams, charts, forms and tappable buttons; The Indian Express adds in-chat tools such as calculators, bill splitters and simple games. OpenAI says the interface appears progressively, assembled from a library of native components and a compiler, and that GPT-6 can begin answering while it is still thinking or using tools. "ChatGPT can now generate a custom UI for you!" Sam Altman wrote on X.

OpenAI's numbers

All three figures below are OpenAI's, from internal evaluations, and none has been reproduced outside the company:

  • Faster starts on search questions. OpenAI says GPT-6 Instant starts answering questions that need web search 44% sooner on average than GPT-5.6 Instant. That is time to the first words, not time to a finished answer.
  • Reasoning at lower latency. On an internal eval of everyday agentic tasks, OpenAI says GPT-6 at Extra High effort began responding in about the time GPT-5.6 took at Medium while scoring higher than GPT-5.6 at Extra High (via The Indian Express). The eval itself has not been published.
  • Safety. OpenAI says GPT-6 is more resistant to multi-turn attempts to get around its safeguards and has stronger protections against cyber and biological misuse.

Independent leaderboards rank GPT-6 Sol and Luna as API models. None of them measures the ChatGPT experience, Intelligent UI, or time to first token inside the app, so the capability claims here remain OpenAI's until someone tests the product.

Limits

  • OpenAI's announcement page and help-centre article could not be loaded when we checked. OpenAI's wording comes from its X post and the copy of the announcement on its developer forum; the eval figures come from The Indian Express's reporting of the post.
  • Speed and quality figures are internal evals with undisclosed methods.
  • Which effort level each plan gets by default in ChatGPT is not confirmed in the sources we read.

Sources

Prior Coverage

Earlier Times of AI reporting on this thread.

Scroll to continue reading