Today, OpenAI announced it is lifting restrictions on text-based conversations for all ChatGPT users, a service that has recently reached over 1 billion weekly users.

This interaction will be powered by the new GPT-5.6 Luna model, which will serve as the default for Free and Go users, taking the place of GPT-5.5.

Users will also receive a new “Think” button, enabling them to choose enhanced reasoning capabilities for intricate questions. OpenAI clarified that separate restrictions will remain in place for file uploads, images, voice inputs, and image generation.

The revision also introduces enhancements for Plus and Pro users, granting them access to an upgraded GPT-5.6 Sol model that is more effective for rapid tasks such as inquiries, online research, offering advice, planning, content creation, and decision-making. The company mentioned that this new model will provide more concise and powerful responses. (It’s worth noting that this is a distinct version from the GPT-5.6 Sol utilized for Codex and Work, which remains unchanged.)

Plus and Pro users will also have a thinking slider, allowing them to adjust the extent of deep thought the model applies to an answer. They can modify the thinking slider based on the complexity and steps required to resolve a query.

According to OpenAI, an internal assessment demonstrated that, in comparison to GPT-5.5-Instant, factual inaccuracies were reported to be 62% less frequent for GPT-5.6 Luna and 68% less for GPT-5.6 Sol.

The revamped version of GPT-5.6 Sol is available to Plus and Pro users as of today, while the modifications for Free and Go users will begin rolling out this week. By next week, users will enjoy unlimited text conversations and have access to the new Think button for more challenging questions.