Anthropic, the artificial intelligence company behind the Claude chatbot, has extended access to a model it calls Claude Mythos to Australia, according to a report from idm.net.au.
What makes this notable is how the model is described. According to idm.net.au, Claude Mythos is a "highly advanced model deemed too dangerous for public release." Rather than opening it to everyone, Anthropic has limited access to a select handful of users — and Australia has now been added to that short list.
The outlet frames the rollout around cyber safety. Its reporting suggests the model may improve cyber safety, but pointedly "not for everyone." In other words, the same capabilities that could help defenders may not benefit all parties equally — a tension that runs through much of today's debate over powerful AI tools.
The available sources are thin on specifics. They do not detail what Claude Mythos can do, who exactly the approved users are, what criteria Anthropic used to grant access, or why the company considers the model too risky for a general launch. Those questions remain open based on the material at hand.
Why it matters: when an AI company labels a model too dangerous to release publicly yet still hands it to chosen users in a particular country, it raises real questions about who gets access to the most capable AI, who decides, and whether such tools make the rest of us safer or less so.