Two weeks after the Trump administration forced Anthropic to pull its most powerful AI models, the government has partially lifted the block with some strings attached.

This Tweet is currently unavailable. It might be loading or has been removed.

The US government has allowed Anthropic to restore access to Claude Mythos 5 for more than 100 American companies and government agencies, according to Semafor, marking a significant de-escalation in a tense standoff between the Trump administration and one of the AI industry's most prominent labs.

Commerce Secretary Howard Lutnick informed Anthropic in a letter Friday that adequate safeguards were in place to permit access for a defined list of trusted partners, Semafor reported. Lutnick also noted that Anthropic has committed to working with the government on future model protocols and releases.

The partial restoration comes roughly two weeks after the administration issued export controls ordering Anthropic to suspend all access to both Mythos 5 and Claude Fable 5 — its publicly available counterpart — for any foreign nationals, including Anthropic's own employees.

In our reporting on the pull-down, the government cited national security concerns without providing specifics. While Anthropic said it believed officials had been shown a method to jailbreak Fable 5, the company characterized the vulnerabilities as relatively minor and noted other models could surface the same issues.

Fable 5 remains offline for now. Semafor reports that people close to the talks say a path toward restoring it is taking shape, but no timeline has been confirmed.