SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is broadening its reach by offering Free and Go users worldwide unlimited text interactions with GPT-5.6 Luna. The company announced this change will start next week, with safeguards against misuse still in place. During this rollout, GPT-5.6 Luna will become the default model for affected accounts. Additionally, free users will gain a Think button, enabling the system to dedicate more time to complex queries before providing a response.
This new access is limited to text conversations and does not lift restrictions on file uploads, image creation, or other features that demand extra processing. OpenAI confirmed that limits for these tools will stay in effect. The Think feature will accompany unlimited messaging next week, offering GPT-5.6 Luna additional reasoning capacity for tasks needing more analysis. Controls for features beyond basic chat, such as specific tools with different usage limits, remain separate.
OpenAI unveiled GPT-5.6 as a family of models, including Luna, Sol, and Terra. Luna is designed as the faster, more affordable option. The company noted ChatGPT now has around one billion weekly users engaging in searches, research, planning, writing, and other activities. With this update, users on Free and Go plans will have Luna as their primary ChatGPT version. The expansion improves access to the core chat service while maintaining tool-specific restrictions.
Enhanced reasoning models now available for paid subscriptions
Subscribers of ChatGPT Plus and Pro are also getting an upgraded version of GPT-5.6 Sol for daily use. Sol offers quick replies and deeper reasoning within the paid experience. OpenAI mentioned the update delivers more targeted answers and minimizes unnecessary details when a simpler response suffices. Paid plans now feature a reasoning slider across web, mobile, and desktop, allowing users to adjust how much processing ChatGPT applies to tasks like research, coding, planning, or writing.
The company also observed a significant drop in factual inaccuracies during internal testing of its latest models. Tests focused on financial, medical, and legal prompts requiring detailed, accurate responses. OpenAI reported factual errors decreased by approximately 62% with GPT-5.6 Luna versus GPT-5.5 Instant, and by roughly 68% with GPT-5.6 Sol. The company added that Sol improved its handling of dates, numbers, sources, rules, and assumptions when accuracy was critical.
Unlimited chat access maintained with separate safety protocols
OpenAI emphasized that its latest ChatGPT models include extra safeguards for users under 18. These measures aim to prevent romantic roleplay, age-restricted challenges, and claims that could substitute real-world relationships. They also set boundaries around sexual content, eating disorders, dangerous activities, restricted products, and graphic violence. The company noted new evaluations have been implemented for responses involving younger users. The unlimited chat feature will continue to operate within existing abuse safeguards when expanded to Free and Go accounts.
This update focuses on the main ChatGPT service rather than all OpenAI products utilizing GPT-5.6 models. OpenAI clarified that GPT-5.6 Sol versions used in ChatGPT Work and Codex will remain unchanged with this release. Paid users can access the updated Sol model and reasoning controls immediately. Free and Go users will see Luna as their default model this week, with unlimited chats and the Think button coming next week. Separate limits will still apply to images, uploads, and other tools.