The U.S. government has given Anthropic the green light for a limited re-release of its Mythos 5 AI model, following weeks of debate over national security risks. CNN reports the move partially rolls back a ban that had blocked the model entirely. Axios notes the administration is close to allowing Anthropic to restore the model, with conditions.
This decision marks an escalation in government control over frontier AI, as officials assert new authority to dictate which companies and users can access advanced models. The ban had stemmed from concerns that Mythos 5's capabilities could be exploited for cyberattacks or other malicious uses. OpenAI has separately agreed to let the administration screen users of its latest model.
According to NBC News, the release is limited, though precise conditions remain unspecified by sources. Multiple outlets confirm the partial green light, with the WSJ reporting the Trump administration rolled back part of the model ban. No specific metrics on performance or user caps were disclosed.
The limited release allows Anthropic to regain some commercial momentum while keeping safeguards in place. Critics argue the government's intervention sets a precedent that could stifle innovation. Future access may depend on ongoing federal review of AI model capabilities and export risks.
The WSJ notes some industry observers worry the partial ban still leaves uncertainty for developers relying on Mythos 5. A spokesperson for Anthropic declined to comment on specific rollout plans.