ChatGPT Update: GPT-5.6 Sol & Luna Enhanced, Safer AI

OpenAI has rolled out a major update for ChatGPT, enhancing the GPT-5.6 Sol model and making GPT-5.6 Luna the default for free users. This update focuses on improving factual accuracy, response quality, and safety across the ChatGPT experience. For Plus and Pro users, there's a new reasoning slider to control how much the model "thinks" before answering. Free and Go users will soon get unlimited GPT-5.6 Luna text chats and a "Think" button for complex questions. GPT-5.6 Sol is refined for everyday conversations, web searches, and various tasks, providing more direct answers and cleaner formatting. It adjusts response length based on question complexity. A key improvement is factual accuracy, with GPT-5.6 Sol better handling dates, numbers, and web information. It's designed to identify and correct incorrect assumptions. Internal evaluations showed a significant reduction in factual mistakes, with GPT-5.6 Sol having 68% fewer errors and GPT-5.6 Luna 62% fewer compared to GPT-5.5 Instant. The update also includes additional safety improvements, particularly for younger users, by avoiding romantic roleplay with minors and stronger safeguards around sensitive content. This means a more reliable and safer ChatGPT experience for everyone.