News in Short
- OpenAI is removing text chat limits for Free ChatGPT users.
- GPT-5.6 Luna becomes the default model for Free and Go users.
- A new “Think” button adds deeper reasoning for complex questions.
- Plus and Pro users get the upgraded GPT-5.6 Sol model.
- File uploads, image generation and voice chats will continue to have separate limits.
OpenAI has announced one of the biggest updates to ChatGPT’s free tier by removing limits on text-based conversations. The change comes shortly after the company revealed that ChatGPT has crossed one billion weekly users, making unlimited text chats available to Free users while introducing new AI models and reasoning controls across subscription tiers.
The rollout begins this week, with unlimited text conversations arriving for Free and Go users next week.
Unlimited text chats arrive for free users
OpenAI is eliminating limits on text-based conversations for Free ChatGPT users, allowing them to continue chatting without running into message caps.
The company says GPT-5.6 Luna will power the free experience, replacing GPT-5.5 as the default model for both Free and Go subscribers.
However, limits on image generation, file uploads, voice conversations and other multimodal features will remain in place.
New “Think” button for tougher questions
Alongside unlimited chats, OpenAI is introducing a new “Think” button for Free and Go users.
The feature allows users to ask ChatGPT to spend more time reasoning through difficult questions before generating a response, making it better suited for complex problems and multi-step tasks.
The company says the feature will begin rolling out next week.
Plus and Pro users get smarter controls
OpenAI is also updating the experience for paid subscribers.
ChatGPT Plus and Pro users now receive an upgraded GPT-5.6 Sol model that is designed for everyday tasks such as writing, planning, web research, advice and decision-making. According to OpenAI, the updated model delivers more concise and reliable answers.
Paid users are also getting a new thinking slider, allowing them to control how much reasoning ChatGPT uses based on the complexity of a task.
OpenAI says accuracy has improved
According to OpenAI’s internal evaluations, GPT-5.6 Luna reduces factual errors by 62 percent compared to GPT-5.5 Instant, while GPT-5.6 Sol cuts factual mistakes by 68 percent.
The updated GPT-5.6 Sol model is available to Plus and Pro subscribers starting today. Meanwhile, GPT-5.6 Luna, the Think button and unlimited text chats will roll out to Free and Go users over the coming week.
