OpenAI has updated GPT-5.6 Sol inside ChatGPT, targeting accuracy and consistency improvements, while simultaneously expanding GPT-5.6 Luna access to free-tier users with no daily chat limits.
The two-model structure matters: Sol and Luna are not the same product. Understanding which model you are actually talking to, and why OpenAI is tiering them this way, is the core story here. Free users now get unlimited everyday conversations via Luna, while Sol serves as the higher-precision option.
The full post details what changed in Sol's behavior and where the accuracy gaps previously were. If you use ChatGPT in any capacity, the model tier distinction and what OpenAI considers 'everyday' versus precision workloads is worth reading in full.
[READ ORIGINAL →]