According to the Financial Times, citing a number of people familiar with the matter, Anthropic refused to submit its latest model to the British AI Security Institute for testing before it was released. Claude Mythos 5.1 is only open to audited US institutions and is not available to the UK AI Security Institute. The report quoted data from the British government and pointed out that the decision raised concerns within Whitehall and the agency. Anthropic declined to comment. The UK Cabinet Office said the AI Safety Institute “is still working closely with industry partners, including Anthropic, to improve the safety of the models.”

According to the Financial Times, citing a number of people familiar with the matter, Anthropic refused to submit its latest model to the British AI Security Institute for testing before it was released. Claude Mythos 5.1 is only open to audited US institutions and is not available to the UK AI Security Institute. The report quoted data from the British government and pointed out that the decision raised concerns within Whitehall and the agency. Anthropic declined to comment. The UK Cabinet Office said the AI Safety Institute “is still working closely with industry partners, including Anthropic, to improve the safety of the models.”