SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is expanding access for Free and Go users by removing chat limits on GPT-5.6 Luna. The company announced that unlimited text conversations will start next week, with abuse safeguards still in place. During this rollout, Luna will be the default model for these plans. At the same time, Free users will see a new Think button. This feature allows the model to spend more time reasoning through complex questions before responding.
The new allowance applies to standard text chats, not all ChatGPT functions. Limits still exist for image generation, file uploads, and other resource-heavy tools. OpenAI said the Think button will arrive alongside unlimited text access next week. Users can activate it when a question needs extra reasoning. This change broadens the capabilities for Free and Go accounts, while keeping separate limits on other tools and services.
GPT-5.6 includes several models such as Luna, Sol, and Terra, each designed for different performance needs. Luna is the faster, more affordable option in the lineup. OpenAI stated that around one billion people now use ChatGPT weekly for research, planning, writing, and search. With this latest update, Luna will be the main ChatGPT model for Free and Go users. Other GPT-5.6 variants remain available within specific products and plans to suit different workloads and access levels.
Free access grows, but tool limits stay in place
ChatGPT Plus and Pro subscribers now have access to an updated GPT-5.6 Sol model for everyday use and more advanced reasoning. The company explained that Sol provides more targeted answers and cuts down on unnecessary detail when shorter responses suffice. Paid users also gain a reasoning slider on web, mobile, and desktop. This control allows adjustments in how much processing the model uses for each request. It’s useful for research, coding, writing, planning, and other tasks requiring different levels of reasoning.
OpenAI also shared results from internal tests focused on accuracy. These tests involved financial, medical, and legal questions needing detailed factual responses. The company reported about 62% fewer answers with at least one factual error using GPT-5.6 Luna compared with GPT-5.5 Instant. GPT-5.6 Sol showed a reduction of about 68% under the same testing method. Improvements were also noted in responses involving dates, numbers, sources, rules, and assumptions when these details influenced the answers.
New models include safeguards for younger users
Safeguards have been added to the GPT-5.6 models for users under 18. These protections cover romantic roleplay, age-restricted challenges, and claims that might encourage users to replace real-world relationships. Additional boundaries prevent content related to sexual material, eating disorders, dangerous activities, restricted goods, and graphic violence. OpenAI said it conducted extra evaluations for responses involving younger users. These abuse safeguards will remain when unlimited chat is available to Free and Go accounts during the expansion of ChatGPT access.
This update applies to the main ChatGPT service only. It does not alter every product using GPT-5.6. Versions of GPT-5.6 Sol operating in ChatGPT Work and Codex stay unchanged. Plus and Pro users can now access the updated Sol model and reasoning controls. Free and Go users will see Luna as their default model this week. Unlimited chats and the Think button will be available next week. Separate limits will still apply to image generation, uploads, and other intensive tools.