U.S. government restricts access to OpenAI's new AI model (www.zeit.de)

OpenAI has announced a restricted access model for its latest AI release, the GPT-5.6 series, in response to a directive from the U.S. government. Only a select group of trusted partners will have early access to these models, initially limited to the U.S., with plans to expand to international partners soon. This move signals a significant shift in governmental oversight of AI technologies, as the U.S. has previously encouraged tech companies to collaborate freely, but now is exercising more control, particularly concerning national security risks posed by advanced AI capabilities. The GPT-5.6 series includes the flagship model Sol, an intermediate model Terra for everyday use, and a budget-friendly variant Luna, with Terra reportedly costing half of its predecessor. This strategic release aligns with OpenAI's competitive stance against rivals like Anthropic, which recently faced restrictions that led to the withdrawal of its AI models from broader access. The government's heightened scrutiny reflects growing concerns over AI's potential to identify security vulnerabilities, prompting companies to work closely with regulators to ensure compliance while striving to innovate in a rapidly evolving landscape.