OpenAI: GPT-5.6 Sol for Paid, Weaker Model for Free Users

OpenAI is updating its ChatGPT models, giving paying users a more focused experience and restricting free users to a weaker version. Plus and Pro subscribers now have an improved GPT-5.6 Sol and a slider to adjust response depth. Here's the thing: OpenAI says GPT-5.6 Sol now cuts unnecessary detail, delivering direct answers for simple questions and more concise responses for complex tasks. They also claim better factual accuracy, with internal evaluations showing a significant drop in errors compared to the previous model. What's interesting is that paying users can now control how much effort ChatGPT puts into a response, with settings for everyday questions up to planning and research. The bottom line for free users is that they will get unlimited text chats with the cheaper GPT-5.6 Luna model later this week. However, this smaller model is generally more error-prone and less reliable. This means free users are losing access to OpenAI's most powerful reasoning capabilities.