OpenAI is making ChatGPT more useful for people who use the service without a paid subscription. Starting next week, Free users will gradually get unlimited text conversations, removing one of the limits that often stopped people from continuing a chat.
The change will also apply to users on the Go plan. Along with unlimited text chats, both plans will move to GPT-5.6 Luna as the standard model instead of GPT-5.5.
OpenAI says GPT-5.6 Luna was designed to handle everyday questions and tasks more effectively. The company is also bringing back a feature called “Think,” which gives users a way to ask ChatGPT to spend more time working through a difficult question before answering.
There is an important catch, though. The unlimited access only covers normal text conversations. Chats that involve image generation or uploaded files will still have usage limits.
For Plus and Pro subscribers, OpenAI has also updated the way users control how much reasoning the chatbot uses. The new system comes with a slider placed near the message box. Users can move it from an instant response setting to a deeper “Pro” level.
The lower setting is aimed at simple questions where users want a quick answer. Higher settings are designed for tasks that need more planning, detailed writing, coding, or research.
OpenAI has also introduced an updated GPT-5.6 Sol model for Plus and Pro users. The company says this version is designed to give more direct answers. It is already available to subscribers on those plans, while the versions used for Work and Codex have not been changed.
Another major focus of the update is accuracy. OpenAI says its internal tests showed fewer factual mistakes in the new models. In tests using questions from areas such as finance, medicine, and law, the number of responses containing at least one error was reduced by 62% with GPT-5.6 Luna and 68% with GPT-5.6 Sol compared with GPT-5.5 Instant.
These numbers come from OpenAI’s own testing, so they should not be treated as an independent measure of accuracy. Still, reducing mistakes is important, especially when people use AI for more serious questions.
OpenAI is also adding stronger protections for users under 18. The company says the new models are trained to refuse certain requests involving romantic roleplay, dangerous online challenges, and conversations that try to replace real human relationships.
The changes come as OpenAI faces growing attention over how AI systems can affect younger users and people in difficult situations. The company has faced lawsuits in the United States related to claims that earlier versions of ChatGPT contributed to harmful real-world situations.
For regular Free and Go users, however, the biggest change is much simpler: text conversations are becoming easier to use without constantly worrying about hitting a message limit. The rollout will happen gradually, so the new access may not appear for every account at the same time.
