OpenAI, the US-based artificial intelligence research lab, has announced the public launch of its most advanced model, GPT-5.6, set for Thursday.The delay in the release followed government requests amid national security concerns about potential misuse of powerful AI systems by military or intelligence agencies in adversarial nations.The US Department of Commerce approved the broad launch after additional testing under Washington’s new oversight framework for frontier AI.Previously, OpenAI had limited access to a small group of vetted partners.The tech firm will also release Terra (mid-tier model) and Luna (cost-efficient option).This follows Anthropic’s recent approval to resume Fable and Mythos models after suspension over security risks.Experts warn that fully securing AI against jailbreaks is nearly impossible, highlighting ongoing challenges in balancing innovation with safety.