China may have accessed Mythos
TL;DR
Semafor reports that White House concerns over a China-linked group possibly accessing Anthropic's Mythos helped drive export restrictions on the model. The White House has not confirmed the China angle, and Trump adviser David Sacks publicly emphasized alleged jailbreak risks instead. If China actually accessed Mythos 5 or Fable 5, the risk is not just use of the model, but distillation: training a weaker model to imitate the stronger one.
Nauti's Take
The uncomfortable part is not the headline alone, but the mix of alleged access, export policy, and shifting public rationale. If a model is supposedly too dangerous for public release but still leaks through Discord groups or outside actors, the safety messaging starts to look brittle.
Anthropic cannot solve that with denials alone. For models at this level, the real questions are who had access, for how long, under what permissions, and whether that access produced a durable advantage.
Briefingshow
This shows how fast frontier models are being treated like chips or cyber capabilities: access itself becomes a geopolitical issue. Even an unconfirmed access claim can shape export controls. For AI labs, model security, access logs, and incident response are no longer just internal security chores; they are becoming national security infrastructure.
Sources