OpenAI has released its new GPT-5.6 model suite, featuring three models: Sol, Terra, and Luna, with improved capabilities in coding, cybersecurity, and biology. The models are priced competitively, with Sol being nearly half the cost of Anthropic’s Claude Fable 5. OpenAI has also emphasized safety and potential misuse, dedicating a significant portion of its announcement to these concerns. The company has implemented a robust safety stack and worked with third-party testers to ensure the model’s security. Why it matters: this release is significant as it showcases OpenAI’s efforts to balance innovation with responsibility, amidst growing regulatory scrutiny in the AI landscape.