White House seeks U.S. review before UK access to new AI models

The White House has asked OpenAI and Anthropic to hold new artificial intelligence models back from the United Kingdom’s AI Safety Institute until U.S. agencies review them, according to reporting that cited administration sources. Anthropic’s Claude Mythos 5.1 was not made available to the institute, while OpenAI’s response to the request was not reported. The request was made by the Office of the National Cyber Director, according to the accounts. The supplied reports do not include a public instruction or direct confirmation from the White House, leaving the request’s precise scope and status unclear. The reports describe it as a call for U.S. review before models are shared with the British institute, not as a requirement for government approval before a model can be released. Anthropic said Claude Mythos 5.1 was available only to a set of U.S. organizations. AISI Director Henry de Zoete said the institute had not received the model. Anthropic also said it was coordinating with Washington on expanding access to domestic and international partners. De Zoete has said AISI retains access to some advanced models and tested OpenAI’s GPT-6 Astra before its release.