Trump Administration Eases Restrictions on Anthropic’s Claude Mythos 5
The US Commerce Department has partially rolled back its ban on Anthropic’s most powerful AI model, allowing select domestic organizations to deploy Claude Mythos 5.
- The US government has authorized Anthropic to restore access to Claude Mythos 5 for over 100 approved domestic corporations and government agencies.
- The consumer-facing version, Claude Fable 5, remains restricted while negotiations between Anthropic and the White House continue over additional safety measures.
- Under the revised terms, approved organizations and Anthropic can now allow foreign national employees to access the Mythos model.
- The original June 12 restriction was triggered by security concerns involving a South Korean telecom firm, alongside jailbreak warnings from Amazon and the NSA.
The Trump administration has partially eased its restrictions on Anthropic’s most advanced artificial intelligence model, Claude Mythos 5. According to a report by Wired, US Commerce Secretary Howard Lutnick notified the AI safety startup that it may restore access to the model for over 100 pre-approved domestic corporations, cybersecurity defenders, infrastructure providers, and government agencies. The decision, first reported by Semafor, comes after weeks of intense negotiations between Anthropic executives and federal officials.
Why it matters
The decision partially reverses a sweeping export control directive issued on June 12, which forced Anthropic to completely disable access to both Mythos 5 and its consumer-oriented counterpart, Claude Fable 5. The White House originally intervened over national security anxieties, particularly after learning Anthropic had granted model access to a South Korean telecommunications company suspected of having ties to China. Additionally, Amazon and the National Security Agency (NSA) had warned administration officials that Fable 5 was vulnerable to jailbreaking, prompting a rapid shutdown of the systems.
In a letter sent to Anthropic co-founder and chief compute officer Tom Brown, Lutnick stated that appropriate safeguards are now in place to allow trusted partners back onto the platform. Under the revised terms, approved organizations and Anthropic itself are also permitted to let their foreign national employees access the Mythos model—a crucial concession, as the initial ban on foreign national access had severely disrupted operations for US-based development teams.
What it means for the AI market
While the new authorization represents progress for Anthropic, the government has stopped short of a full public release. The consumer-facing Claude Fable 5 remains offline as talks between Anthropic and the White House continue. Eduardo Maia Silva, a spokesperson for Anthropic, confirmed that the company is actively working to provision approved cyber defenders and restore access as quickly as possible, while hoping to establish a lasting policy framework for future model releases.
This regulatory friction highlights the growing government oversight of frontier AI models. For enterprises choosing between top-tier LLMs, navigating these sudden regulatory shutdowns is becoming a critical operational risk. Those evaluating how these high-security models perform relative to unrestricted commercial options can read our head-to-head ChatGPT vs Claude comparison to see how the two ecosystems stack up under real-world workloads.
Frequently asked questions
What is Claude Mythos 5?
Claude Mythos 5 is Anthropic’s strongest, high-security AI model, specifically optimized for cybersecurity defense and critical infrastructure protection.
Why did the US government restrict Anthropic’s models?
The Trump administration blocked access on June 12 due to export control concerns regarding foreign nationals, a South Korean telecom firm with suspected Chinese ties, and jailbreak warnings from the NSA and Amazon.
Is Claude Fable 5 available to the public yet?
No. While Mythos 5 has been cleared for select organizations, the consumer-facing Fable 5 remains restricted while Anthropic and the White House negotiate additional safeguards.
Read our hands-on comparison of ChatGPT vs Claude to see how the leading AI models perform under real-world testing.
ChatGPT vs Claude (2026): Which Is Better? (Tested) →Source: Wired. Published June 27, 2026.
Ali has hands-on tested 50+ AI tools and tracks model releases daily. Every verdict here comes from real, paid usage — never vendor demos or sponsored placements.
AI Tools Worth is independent and unsponsored. Some linked guides contain affiliate links — they never change our verdicts.