OpenAI is making one of ChatGPT’s most consequential consumer changes yet: text conversations are set to become unlimited for Free and Go users. The updated free experience will run on GPT-5.6 Luna, while a new Think button will offer extra reasoning for more difficult questions.
The announcement lowers a familiar barrier for people who rely on ChatGPT for study, planning, writing and everyday problem-solving. However, “unlimited” applies specifically to text chats. Separate usage limits will continue for features such as files, images, voice and other tools.
Background: why ChatGPT’s free limits mattered
Free access has helped turn ChatGPT into a mainstream tool, but usage caps could interrupt a conversation or move users to a less capable experience after a threshold. That mattered most for tasks that benefit from iteration: refining a résumé, debugging code, comparing options or working through a lesson step by step.
OpenAI’s new model family uses capability-tier names alongside the generation number. GPT-5.6 Sol is the flagship tier, while Luna is designed to be the fastest and most affordable option. For Free and Go accounts, Luna is replacing GPT-5.5 as the default ChatGPT model.
What OpenAI announced
Unlimited text conversations for Free and Go
OpenAI says it is removing text-based rate limits for ChatGPT Free and Go users. GPT-5.6 Luna is becoming their default model during the current rollout, with unlimited text chats scheduled to arrive the following week after the 6 August announcement.
That timing distinction is important. Users may see Luna before the unlimited-chat policy or Think button appears on their account, because the features are rolling out in stages rather than switching on everywhere at once.
A Think button for harder questions
Free and Go users will also receive a Think button. Selecting it gives GPT-5.6 Luna more time to work through a request that needs additional reasoning. In practice, it should be most useful for multi-step questions, structured comparisons, logic problems and tasks where a quick first response is not enough.
The feature does not guarantee a correct answer. It is better viewed as a way to allocate more reasoning effort when the task justifies it.
GPT-5.6 Sol changes for paid users
Plus and Pro subscribers are receiving an updated version of GPT-5.6 Sol for ChatGPT, along with a slider that controls how much thinking the model applies. OpenAI describes the update as more factual, focused and compact, especially for everyday questions, research, planning, advice and writing.
According to OpenAI’s internal evaluation, factual errors were 62% less common with GPT-5.6 Luna and 68% less common with GPT-5.6 Sol than with GPT-5.5 Instant. Those are company-reported results, not a guarantee for every prompt. OpenAI also says the ChatGPT update does not change the Sol versions used by Codex and ChatGPT Work.
Why ChatGPT unlimited free chats matter
The immediate benefit is continuity. A student can keep asking follow-up questions, a jobseeker can revise several versions of an application, and a small business owner can continue planning a campaign without worrying that ordinary text messages will exhaust a quota.
The change may also intensify competition among consumer AI assistants. Text chat is the entry point for most users, and removing the basic cap makes it easier to build a daily habit around one service. Competitors will face more pressure to differentiate through model quality, integrations, privacy, specialist tools or generous free access of their own.
Practical impact for users, businesses and creators
For individuals, the best use cases remain interactive ones. Instead of requesting a finished answer in one prompt, users can ask for an outline, challenge assumptions, request examples and improve the result over several turns. Unlimited text makes that workflow more accessible.
Creators and small teams can use the free tier for brainstorming, editing, summaries and basic research preparation. Developers can discuss code, error messages and architecture ideas in text, although file uploads and advanced tools may still hit separate limits. Businesses should also remember that free consumer access is not a substitute for an approved workplace AI policy, suitable data controls or enterprise terms.
Risks, limitations and concerns
The biggest caveat is the scope of “unlimited.” It does not mean unrestricted access to every ChatGPT capability. OpenAI says separate limits remain for files, images, voice and image generation. Abuse guardrails and service protections can also apply, so automated scraping or prohibited activity should not be assumed to be covered.
More access does not eliminate hallucinations. Users should verify dates, prices, quotations, legal or medical information, and other high-stakes claims against primary sources. The reported reduction in factual errors is encouraging, but even a stronger model can misunderstand context or provide a confident answer that is wrong.
Privacy is another consideration. Avoid pasting passwords, confidential client records, unpublished business information or sensitive personal data into a consumer chatbot. Organisations should define what employees may share and use managed services when stronger governance is required.
What to watch next
The first test will be rollout consistency: whether Free and Go users receive Luna, Think and unlimited text on schedule across web and mobile apps. It will also be worth watching whether OpenAI publishes clearer fair-use details or adjusts remaining tool limits as demand grows.
Model quality will matter as much as access. Users should look for better source handling, fewer factual mistakes and a useful balance between concise answers and deeper reasoning. Paid plans will need to justify themselves through stronger models, larger tool allowances and workflow features rather than basic text availability alone.
Conclusion
ChatGPT unlimited free chats could make sustained AI assistance available to a much wider audience. GPT-5.6 Luna as the default model and the new Think button should improve the free experience, particularly for iterative work and harder questions. The headline still needs careful reading, though: unlimited text is not unlimited files, images, voice or tools, and users must continue to verify important outputs.