Calvin Wankhede / Android Authority
TL;DR
- OpenAI’s latest ChatGPT update brings meaningful upgrades to the GPT-5.6 lineup.
- The top-tier GPT-5.6 Sol model will now make fewer factual errors, limit unnecessary information, and can adapt its reasoning based on your query.
- Free users are moving from GPT-5.5 Instant to GPT-5.6 Luna with unlimited text chats and a new Think button for deeper reasoning.
ChatGPT’s latest 5.6 model lineup made its official debut early in July. Less than a month later, the AI giant is already giving the state-of-the-art models a brain and accessibility boost.
As you may recall, the 5.6 series consists of three models: Sol, Terra, and Luna. There’s no update to Terra, but GPT-5.6 Sol and GPT-5.6 Luna are getting meaningful upgrades. These range from making the models more reliable, accurate, and focused, to expanding access for users who don’t want to pay for top-tier AI.
The biggest change is that GPT-5.6 Luna will soon be the default ChatGPT model for all non-paying users. This change is set to kick in sometime this week, making the speed-focused lightweight model more accessible globally.
Elsewhere, starting next week, non-paying users will also gain access to unlimited text chats with GPT-5.6 Luna, complete with a new “Think” button for more mentally demanding queries. This will prompt the model to take more time to work through your query and generate an answer.
Essentially, free users will go from GPT-5.5 Instant to the more capable GPT-5.6 Luna with unlimited chats and an option for deeper thinking and reasoning as an added bonus.
Within the same update, OpenAI also gives GPT‑5.6 Sol, the heavy hitter of the 5.6 lineup, a series of refinements. Limited to Plus and Pro users, the model now “delivers more focused answers, adapts its level of detail to the question, avoids unnecessary formatting, and offers a helpful correction when simply agreeing wouldn’t be useful.”
For reference, a prompt like “Can I bike from the Mission to Ocean Beach after work today without getting soaked?” will attempt to address the main answer first, followed by key details that actually matter. The model provides a direct answer while avoiding extra detail when it does not help.
Beyond that, the model is trained to make fewer mistakes, especially when it comes to dates, numbers, sources, rules, or assumptions. The AI giant says that factual errors with GPT‑5.6 Sol are 68% less common than they are with GPT-5.5 Instant. Regardless, users should still avoid blindly accepting AI-generated information as fact. Although today’s update makes ChatGPT overall more reliable, hallucination concerns are still very real. This makes cross-checking important information imperative.
Lastly, within the same update, OpenAI is also giving Plus and Pro users a seamless way to specify how much reasoning effort ChatGPT should put into an answer.
The slider, as seen in the short GIF above, is accessible through a new speedometer icon within the app’s toolbar. From there, you drag to toggle between Instant, Medium, High, Extra High, and Pro level thinking.
The slider and the updated version of GPT-5.6 Sol are available to Plus and Pro users starting today. Free access to GPT-5.6 Luna can be expected this week, while the new Think button will land next week.
Thank you for being part of our community. Read our Comment Policy before posting.
