Anthropic Regains Access to Claude Models After Federal Review

Anthropic announced Tuesday that federal restrictions on its Claude Fable 5 and Mythos 5 AI models have been lifted, ending a dispute that forced the company to disable public access for more than two weeks.

The AI company said it will begin restoring access Wednesday, according to details first reported by CBS News. Commerce Secretary Howard Lutnick confirmed his team worked with Anthropic to analyze and approve Fable 5, aiming to strengthen American leadership in artificial intelligence.

The dispute over jailbreaks and export controls

Anthropicintroduced Claude Fable 5 to the public earlier this month with safeguards designed to prevent misuse for cyberattacks. The company planned to release Mythos 5, a version with fewer guardrails, to select major companies for testing.

Days after launch, the federal government imposed export controls requiring Anthropic to block access by foreign nationals—a directive so broad the company said it effectively required disabling the models entirely.

The government's concerns centered on a potential jailbreak technique that could bypass the AI's safety guardrails. Anthropic disputed the severity, arguing the vulnerabilities were relatively simple and could be replicated with other AI models on the market. The company warned that applying such standards industry-wide would halt new model deployments from all frontier AI providers.

Last week, the Commerce Department began allowing certain companies and government entities to use Mythos 5 again, according to reports from CNBC and The Wall Street Journal, signaling a partial easing before Tuesday's full restoration.

Why it matters

The episode illustrates the Trump administration's struggle to balance AI safety concerns with its stated goal of avoiding regulation that could hamper U.S. competitiveness against China. The resolution suggests the government is developing case-by-case approaches rather than blanket restrictions, though the criteria for approval remain unclear. For AI companies, the incident demonstrates that even voluntary cooperation with federal reviews can trigger operational disruptions.

Broader tensions between Anthropic and the administration

This isn't Anthropic's first clash with the Trump administration. Earlier this year, the company pushed for formal guardrails preventing the Pentagon from using Claude models for autonomous weapons or mass surveillance of Americans. When negotiations failed, President Trump ordered federal agencies to stop using Anthropic's technology, and Defense Secretary Pete Hegseth labeled it a supply chain risk.

Anthropic sued, and a federal judge blocked those restrictions, calling them "Orwellian" and an effort to cripple the company. The government is appealing that decision.

President Trump signed an executive order this month establishing a voluntary 30-day review process for advanced private AI models in collaboration with the federal government. White House chief of staff Susie Wiles thanked companies Tuesday for cooperating with the order, including on advanced model access and guardrail testing.

CBS News first reported these developments.