OpenAI has officially launched a limited preview of its GPT-5.6 model family, introducing three distinct variants: Sol, Terra, and Luna. This new generation of large language models aims to provide developers and organizations with greater flexibility in balancing performance, efficiency, and cost.

GPT-5.6 Sol is the flagship model, designed for demanding applications such as software engineering, complex computer use, professional knowledge work, scientific research, and cybersecurity. Its capabilities are enhanced by new features like 'max reasoning effort' and 'ultra mode', enabling deeper analysis and sub-agent utilization in complex tasks.

Complementing Sol, GPT-5.6 Terra delivers performance comparable to the previous GPT-5.5 model but at a significantly reduced cost. Luna, the third variant, is positioned as the most cost-efficient model, offering advanced AI capabilities at OpenAI's lowest price point.

Initially, the GPT-5.6 models will be available through the API and Codex to a select group of trusted partners and organizations. A broader public release to users of ChatGPT, Codex, and the API is planned for the coming weeks. The tiered naming system allows for independent evolution of these model categories.