SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is broadening access to ChatGPT by offering Free and Go users unlimited text-based interactions with GPT-5.6 Luna. The company announced that this change will take effect next week and will be implemented with safeguards to prevent misuse. During the current rollout, GPT-5.6 Luna will also become the default model for those accounts. Additionally, free users will now see a Think button that allows the system to dedicate more time to tackling complex questions before providing a response.
This new feature applies solely to text conversations and does not eliminate all usage restrictions within ChatGPT. OpenAI emphasized that limits will still be in place for file uploads, image generation, and other tools that demand extra processing power. The Think feature will launch alongside unlimited messaging next week, offering GPT-5.6 Luna additional reasoning time for tasks requiring deeper analysis. The company maintains separate controls for features beyond standard text interactions, including tools accessible under different usage allowances.
GPT-5.6 has been introduced by OpenAI as a family of models, which includes Luna, Sol, and Terra. Among them, Luna is designed to be the faster and more cost-effective option. According to the company, ChatGPT now serves roughly one billion users each week across tasks like searches, research, planning, writing, and more. With this latest update, users on Free and Go plans will be provided Luna as their primary ChatGPT model. The rollout enhances access to the core chat service while maintaining limits on specific tools.
Updated reasoning model for paid subscriptions
Subscribers of ChatGPT Plus and Pro will also benefit from an upgraded version of GPT-5.6 Sol for daily use. Sol offers both quick replies and more in-depth reasoning within the paid environment. OpenAI stated that this update results in more targeted answers and minimizes unnecessary details when a simple response suffices. Paid users will also gain access to a reasoning slider across web, mobile, and desktop platforms, enabling them to adjust the amount of processing power ChatGPT employs for research, planning, coding, writing, and other tasks.
In internal assessments, OpenAI observed a decline in factual inaccuracies in its latest models. These tests, which involved financial, medical, and legal prompts requiring detailed factual responses, showed that answers with at least one factual mistake decreased by approximately 62% with GPT-5.6 Luna compared to GPT-5.5 Instant. For GPT-5.6 Sol, the reduction was roughly 68%. The company also noted improvements in Sol’s ability to handle dates, numbers, sources, rules, and assumptions when precision was critical.
Separate safety measures for unlimited messaging
OpenAI stated that its newest ChatGPT models incorporate additional safeguards, especially for users believed to be under 18. These measures are designed to prevent romantic roleplay, age-restricted challenges, and claims that could substitute genuine relationships. Boundaries addressing sexual content, eating disorders, dangerous activities, restricted items, and graphic violence have also been introduced. The company highlighted that new evaluations have been added for responses involving younger users. When expanded access reaches Free and Go accounts, unlimited text conversations will still operate within the same safeguards against abuse.
This update mainly focuses on the primary ChatGPT experience and does not affect other OpenAI products utilizing GPT-5.6 models, such as ChatGPT Work and Codex, which remain unchanged. Currently, paid users can access the updated Sol model along with its reasoning controls. For free and Go users, Luna will serve as the default model beginning this week. The rollout of unlimited text chats and the Think button will follow next week, with other limitations for images, uploads, and additional tools remaining in place.