OpenAI says access to its new GPT-5.6 model is limited at the request of the US government (www.businessinsider.com)

OpenAI has announced that access to its new GPT-5.6 models, including variants named Sol, Terra, and Luna, will initially be limited at the request of the US government. This decision follows a preview of the models to government officials ahead of their scheduled launch on Friday. OpenAI is complying with an agreement established with the Department of Defense, which allows the Pentagon to utilize its AI models. The company emphasizes that this temporary restriction should not become the norm, as it can hinder access for users, developers, and cyber defenders who need these advanced tools. This move is significant in the AI/ML landscape as it highlights ongoing government concerns regarding security and responsible AI deployment. OpenAI asserts that by taking this initial step, it aims to foster broader accessibility in the coming weeks while collaborating with the government to establish a sustainable framework for future model releases. The context of limited access echoes similar actions from other companies, such as Anthropic, which has also restricted access to its latest models due to governmental security concerns, underscoring the ongoing dialogue between AI firms and regulatory bodies.