Key Takeaways
- Federal export restrictions on Anthropic’s Mythos 5 and Fable 5 were removed on June 30
- The original ban began June 12 after security researchers identified a critical jailbreak vulnerability
- Anthropic blocked all users globally due to inability to authenticate user nationalities instantly
- The company committed to intelligence sharing and providing early model access to federal agencies
- Growing competition from Chinese AI systems like Fugu and Tulongfeng influenced the policy reversal
Anthropic has resumed worldwide availability of its flagship AI systems following the federal government’s decision to withdraw export restrictions that blocked the models throughout most of June.
Fable 5 became accessible internationally on July 1 through all Anthropic distribution channels. Mythos 5, built on identical architecture but featuring reduced safety limitations, is now available to a select group of American institutions granted federal clearance on June 26.
The Origin of the Restrictions
The prohibition commenced June 12 when federal authorities designated both systems as export-controlled technologies. This classification prevented distribution to international users without explicit authorization.
A cybersecurity disclosure prompted the action. Researchers at Amazon uncovered a jailbreak techniqueâa method for circumventing built-in protectionsâthat enabled Fable 5 to detect software security flaws and generate exploit code demonstrating potential attacks.
With the directive effective immediately and no mechanism to authenticate user citizenship on demand, Anthropic disabled access universally to avoid potential violations.
Security professionals questioned the necessity of the measures from the outset. They pointed out that Anthropic had already committed to implementing comparable security safeguards voluntarily, well ahead of the export regulation. Some observers viewed the prohibition as a negotiating tactic rather than a substantive security intervention, potentially connected to Anthropic leadership’s public objections to the Trump administration’s AI policies.
Terms of the Agreement
Commerce Secretary Howard Lutnick announced that Anthropic committed to actively identifying and mitigating security threats, collaborating with authorities on deployment protocols for upcoming models, and reporting suspicious or malicious usage patterns.
The partnership with federal agencies has expanded significantly. Anthropic will provide designated government entities with advance access to cutting-edge models prior to public launch and share actionable threat data.
Together with Amazon, Microsoft, and Google, the company is developing a standardized system to assess jailbreak severity levels. These commitments align with a June 2 executive directive focused on AI security infrastructure.
International Market Dynamics
Chinese technology firms have started deploying AI systems with performance characteristics comparable to Mythos-level capabilities, notably Fugu and Tulongfeng. This competitive dynamic created challenges for U.S. policymakers attempting to rationalize restricting American AI technologies from international markets.
Lutnick also authorized Mythos distribution to specific White House-vetted clients last week. OpenAI’s most recent systems followed a comparable rollout strategy, launching to approved partners rather than the broader public.
While Anthropic remains a private company, speculators have traded on its estimated valuation through pre-IPO derivative instruments on Hyperliquid’s decentralized exchange. That contract declined approximately 3.7% during the June access suspension.
The Trump administration’s evolving approach to AI regulation has created uncertainty across the sector regarding governance frameworks for future model deployments.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
