OpenAI globally launched its new series of smart models, GPT-5.6, following the completion of security reviews imposed by the US government to ensure that the models comply with national security standards.

The series consists of three main models:

Sol: The most powerful and most advanced, and is intended for complex tasks such as research and programming.
Terra: A balanced choice that combines high performance with low operational cost.
Luna: The fastest and most economical model for everyday tasks and conversations.

The new series features superior capabilities in discovering software vulnerabilities, deeper “thinking” skills to increase the accuracy of results, in addition to incorporating the strongest protection system in the company’s history to ensure adherence to ethical and security standards.

This launch comes as a culmination of extensive coordination between the company and US regulatory authorities, to ensure that technological leadership is balanced with cybersecurity requirements.