OpenAI is enhancing ChatGPT with a more intelligent version of its model, GPT-5.6 Sol, and removing chat limits for free users.

The updated GPT-5.6 Sol is specifically tuned for everyday conversations, aiming to provide more focused answers, adapt detail levels, and avoid unnecessary formatting. It promises direct responses for quick questions and more comprehensive ones for complex tasks like research and planning. The update unifies the 'Instant' and 'Thinking' experiences for Plus and Pro subscribers, who can now use a slider to control the model's reasoning depth, from quick answers to in-depth analysis for writing, coding, and decision-making.

OpenAI also reports that the revised model is more reliable, with factual errors reportedly 68% less common in internal evaluations. Free users will now have unlimited text chats using GPT-5.6 Luna and a new 'Think' button for more complex questions requiring deeper reasoning. This update focuses on regular ChatGPT conversations, with GPT-5.6 Sol powering ChatGPT Work and Codex remaining unchanged.