According to CNBC, the Commerce Department gave Anthropic permission to release Mythos 5 after determining that certain safeguards were in place. “I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model,” U.S. Commerce Secretary Howard Lutnick wrote in a letter viewed by CNBC.
The approval does not cover Fable 5, another advanced Anthropic model that remains blocked. Anthropic had earlier disabled access to both Fable 5 and Mythos 5 after a government export-control directive cited “national security authorities.” The company said the order required it to suspend access for any foreign national, including foreign national Anthropic employees.
AI Model Access Becomes a Policy Risk
The decision shows how closely Washington is now watching frontier AI models. Anthropic had promoted Fable 5 and Mythos 5 as state-of-the-art systems across several benchmarks, with Fable 5 standing out as one of its most advanced public releases.
The move also comes as OpenAI, a privately held AI company backed by Microsoft Corporation (MSFT), announced three new models and said it would first limit access to a small group of trusted partners at the government’s request. OpenAI said it still aims to make the models broadly available in the coming weeks.
