SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is expanding access to ChatGPT for Free and Go plans by removing text chat limits under GPT-5.6 Luna. The company announced that unlimited text conversations will be available starting next week, while safeguards against misuse remain active. During this rollout, Luna is becoming the default model for these plans. Simultaneously, Free users will receive a Think button, enabling the model to spend more time reasoning through complex questions before providing a response.
This increased allowance applies specifically to standard text exchanges and does not extend to all features within ChatGPT. Limits will still be enforced on functions like image creation, file uploads, and other tools that demand additional computational resources. OpenAI indicated that the Think button will be introduced alongside unlimited text access next week, allowing users to activate it when a question warrants extra reasoning time. This change broadens the core conversational capabilities for Free and Go accounts while maintaining separate restrictions for other tools and services.
GPT-5.6 is composed of multiple models, including Luna, Sol, and Terra, each tailored for different performance requirements. Luna is recognized as the faster and more cost-effective option within this lineup. OpenAI stated that approximately one billion people now use ChatGPT weekly for a variety of tasks such as research, planning, writing, and searching. With this update, Luna becomes the primary ChatGPT model for Free and Go users, while other GPT-5.6 variants remain available for different workloads and access levels within various products and plans.
Free plans see expanded access, but tool restrictions stay in place
Subscribers to ChatGPT Plus and Pro are now receiving an updated GPT-5.6 Sol model for everyday use and more intricate reasoning. The company explained that Sol now delivers more targeted answers and minimizes unnecessary detail when shorter responses suffice. Paid plan members also gain a reasoning slider accessible via web, mobile, and desktop platforms. This feature allows users to control how much processing power the model uses for each request, suitable for research, coding, writing, planning, and other tasks requiring different levels of reasoning.
OpenAI shared results from internal tests focused on the factual correctness of its latest models. These evaluations involved financial, medical, and legal questions demanding detailed factual responses. The company reported a reduction of approximately 62% in answers containing at least one factual error with GPT-5.6 Luna compared to GPT-5.5 Instant. The GPT-5.6 Sol model demonstrated an even greater decline of about 68% in errors under the same testing conditions. OpenAI also noted improved accuracy in responses involving dates, numbers, sources, rules, and assumptions when those details impacted the outcome.
New models incorporate safeguards to protect younger users
The GPT-5.6 models feature enhanced safeguards aimed at users under 18, as determined by the system. These protections restrict certain interactions such as romantic roleplay, challenges with age restrictions, and claims that might encourage users to replace real-world relationships. Additional boundaries are enforced against material related to sexual content, eating disorders, risky activities, restricted goods, and graphic violence. OpenAI stated that it has introduced further evaluations for responses involving minors. The same abuse safeguards will continue to be enforced for unlimited text conversations on Free and Go accounts once access is expanded.
This rollout applies to the core ChatGPT service and does not affect every product utilizing GPT-5.6. Versions of GPT-5.6 Sol used within ChatGPT Work and Codex are unchanged by this update. Currently, Plus and Pro users can access the updated Sol model along with reasoning controls. Meanwhile, Free and Go users will see Luna as their default model this week. The availability of unlimited text conversations and the Think button will follow next week, with separate limits remaining for images, uploads, and other intensive tools.