OpenAI is updating ChatGPT’s paid and free tiers with revised GPT-5.6 models, new reasoning controls and broader access to text chats. OpenAI announces in an official blog post that Plus and Pro subscribers can now use an updated GPT-5.6 Sol model in ChatGPT, while free users will move to GPT-5.6 Luna as their default model.
The company says the revised Sol model gives more direct answers, uses less unnecessary formatting and handles factual details more reliably. It is intended to adjust the depth of an answer to the task, from brief everyday questions to multi-step research, planning and writing.
For paid users, a new slider on web, mobile and desktop lets people select how much reasoning effort ChatGPT should apply. The lower setting targets fast responses. Higher settings are aimed at more complex work, including coding and decisions that require additional analysis. OpenAI says Sol will now provide both quick and deeper responses in the Chat experience, rather than presenting separate models with different styles.
What changes for free users
GPT-5.6 Luna will become the default model for Free and Go users during the week of the announcement. From the following week, those users will receive unlimited text chats, subject to abuse controls. A new Think button will let them request more reasoning time for difficult questions.
Limits remain for uploads, image generation and other tools. OpenAI also says the ChatGPT version of Sol differs from the version used in its Work and Codex products, which are not part of this update.
In an internal test covering financial, medical and legal prompts, OpenAI reports that answers with at least one factual error were 62 percent less frequent for Luna and 68 percent less frequent for Sol than for GPT-5.5 Instant. The company did not provide full details of the evaluation in the announcement.
Stay up to date
AI for content creation: the latest tools, tips and trends. Every two weeks in your inbox:
