# OpenAI makes ChatGPT's latest model more accessible to free users GPT-5.6 Luna will become the default model for free users with unlimited chats and new Think feature. - **Source:** [Android Authority](https://brifing.si/source/01374d39-7ab2-4795-8e0a-444a0e7b4327) - **Published:** 2026-08-06 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/ae0469ed-4a90-4b5e-8880-5d1a3319e886 ## Major upgrade for free users OpenAI announced that free users will transition from GPT-5.5 Instant to GPT-5.6 Luna. The update, rolling out this week, will make Luna the default model for all free users. Luna offers a lighter, faster experience along with unlimited text chats. The new 'Think' button allows users to request deeper analysis for complex queries, significantly enhancing the free tier's capabilities. ## Premium model enhancements GPT-5.6 Sol received significant improvements for Plus and Pro users. The model now delivers more focused answers, adjusts detail level based on the question, and avoids unnecessary formatting. It also provides helpful corrections instead of simple confirmations. According to OpenAI, GPT-5.6 Sol makes 68% fewer factual errors compared to GPT-5.5 Instant. However, users are still advised to verify AI-generated information, especially for critical matters. ## Reduced errors in dates and numbers With the update, GPT-5.6 Sol makes fewer mistakes particularly with dates, numbers, sources, rules, and assumptions. For example, when asked 'Can I bike from Mission to Ocean Beach after work without getting soaked?', the model first provides a direct answer followed by key relevant details. While these improvements increase reliability, OpenAI cautions that hallucination risks remain, requiring continued user vigilance. ## New reasoning level adjustment Plus and Pro users can now specify how much reasoning effort ChatGPT should apply to answers. Accessed through a new speedometer icon in the app's toolbar, the slider lets users toggle between Instant, Medium, High, Extra High, and Pro thinking levels. This feature enables optimization of processing power based on query complexity, offering better control over response quality and speed. ## Deep analysis with Think button The 'Think' button, coming next week for free users, prompts the model to spend more time analyzing queries thoroughly. This feature proves particularly valuable for complex problems, technical questions, and multi-step logical reasoning. OpenAI emphasizes that this button represents a significant advantage for free users and advances AI accessibility by bringing premium-grade analysis capabilities to the free tier. ## User experience improvements The update brings numerous subtle but impactful changes to GPT-5.6 Sol that enhance user experience. The model now avoids information overload, focusing directly on answering questions. It also proactively corrects user misconceptions. OpenAI notes these changes will especially benefit educational and research applications where precision and clarity matter most. ## Increased reliability and limitations OpenAI's data shows GPT-5.6 Sol makes 68% fewer factual errors than its predecessor. However, the company warns users to always cross-check critical information as AI hallucinations remain a concern. This caution applies particularly to medical, legal, or financial matters where incorrect information could have serious consequences, highlighting the continued need for human verification. ## New era in AI accessibility This update represents one of OpenAI's most comprehensive improvements for free users. Making GPT-5.6 Luna the default model with unlimited chats provides millions with access to more advanced AI. The Think button bridges the gap between free and premium users by offering deep analysis previously reserved for subscribers, marking a significant step in AI democratization and making sophisticated tools available to wider audiences. ## Sources covering this story - https://www.androidauthority.com/gpt-5-6-update-3695577/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._