US Lifts Export Controls on Anthropic's AI Models After 18-Day Shutdown
The U.S. Department of Commerce has lifted export controls on Anthropic's AI models, Claude Fable 5 and Mythos 5, following an 18-day global suspension. The restoration followed Anthropic's implementation of a classifier to block high-risk techniques.
Key Takeaways
The lifting of export controls followed Anthropic's implementation of a classifier.
The 18-day suspension was unprecedented, driven by national security concerns.
Claude Fable 5 leads in coding performance.
Mythos 5 remains restricted through the 'Glasswing' program.
This situation sets a precedent for government oversight of AI.
1US Lifts Export Controls: A Deep Dive Analysis
On June 30, 2026, the U.S. Department of Commerce lifted export controls on Anthropic's most advanced AI models, Claude Fable 5 and Mythos 5. This decision marked the end of an unprecedented 18-day global suspension, mandated due to national security concerns.
The initial suspension occurred after Amazon researchers discovered that specific prompts could bypass safety guardrails, potentially enabling the generation of cyberattack-related information.
Suspension Context
- 📊 Claude Fable 5 is Anthropic's most powerful model in coding performance, scoring 80.3% on SWE-Bench Pro.
- 🔑 Mythos 5 was restricted to specific U.S. organizations through the 'Glasswing' program.
- 🎯 The restrictions drew criticism from tech executives who argued they gave Chinese open-source developers a competitive edge.
Negotiations and Solutions
Intensive negotiations, led by Anthropic co-founder Tom Brown, resulted in an agreement with the Trump administration to restore access. Anthropic implemented a new classifier blocking high-risk techniques in over 99% of cases.
"Appropriate safeguards have been established," confirmed Commerce Secretary Howard Lutnick.
Implications and Future of AI
- âš¡ This suspension sets a significant precedent for government oversight of frontier AI models.
- 💡 The restoration includes access across platforms like Claude.ai, Claude Code, AWS, Google Cloud, and Microsoft Foundry, though with usage limits until July 7, 2026.
- 📊 Anthropic recently surpassed OpenAI in annualized revenue, reaching $47 billion.