After two weeks of intense negotiations with the U.S. government, Anthropic's Mythos 5 is back in limited operation, but only for a select group of users. According to a letter from Commerce Secretary Howard Lutnick, the cybersecurity model can now be redeployed to certain cyber defenders and infrastructure providers.

Fable 5, however, remains on hold, much like its predecessor. The U.S. government has not lifted the export control directive which barred any foreign nationals – including Anthropic’s own employees – from accessing either model. This decision was made in light of recent work by Anthropic with the U.S. government to address risks associated with both models.

Anthropic spokesperson Danielle Ghiglieri confirmed the news, stating that the company is working quickly to restore access for approved providers and is optimistic about expanding this access further. The move mirrors OpenAI’s recent actions on their GPT-5.6 model, indicating a limited preview release for select entities.

This decision comes under pressure from both within and outside the U.S. AI industry, with fears of falling behind China and the need to ensure top government departments like the NSA still have access. The hope is that this temporary measure will pave the way for broader availability in the coming weeks as a more repeatable process is developed.