US Lifts Partial Export Block on Anthropic’s Mythos 5 AI Model for Cyber Defenders

Key Takeaways

1. On June 27, 2026, the US government authorised a limited release of Anthropic’s Mythos 5 AI model. Only a select group of cyber defenders and critical infrastructure providers can now access it.
    
2. Earlier in June 2026, the US Commerce Department imposed a full export ban. Anthropic had disabled both Mythos 5 and Fable 5 for all customers to comply with the order.

What Triggered the US Government’s Partial Reversal on Anthropic's Mythos 5 Export Ban?

The US Commerce Department, led by Secretary Howard Lutnick, informed Anthropic on June 27, 2026, that “appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.” Anthropic confirmed, “We received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers.”

The original export control order, sent to Anthropic on June 12, 2026, required an approved export licence from the Bureau of Industry and Security (BIS) for any foreign persons to access both Fable 5 and Mythos 5. Anthropic had stated at the time, “The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.” 

How does this Development affect India’s AI and Fintech Sector?

India’s financial sector is a significant user of advanced AI tools for credit decisions and loan management. As of mid-2026, India's fintech platforms have disbursed over 11 crore loans cumulatively, with total loan value crossing ₹2.9 lakh crore, a 45% year-on-year surge. Access restrictions on frontier AI models like Mythos 5 directly affect Indian fintech firms that integrate US-based AI APIs.

According to a September 2025 LoansJagat report on AI in Indian banking, private banks have begun deploying AI avatars and voicebots for loan recovery, with public sector banks evaluating similar tools. Any future tightening of US AI export rules could slow India’s access to frontier models powering these systems. The partial restoration of Mythos 5 access on June 27 is a positive signal, but Fable 5, the more widely used public model, remains restricted.

What are Experts Saying? What is the Path Forward?

The Associated Press reported that Mythos 5 identified weaknesses in “sensitive and secure US government computer systems during a testing exercise,” with an official stating. “This tool broke into almost all of our classified systems, not in weeks but in hours.” This finding triggered the initial June 12 ban.

Maximilian Kasy, an economist and machine learning specialist at Oxford University, noted that large language models like Fable exhibit performance that surpasses expectations, making governance frameworks essential. 

Kasy stated that reliance on guardrails alone is futile given their susceptibility to circumvention, and that what is needed is a global, participatory governance framework rooted in mutual trust. Anthropic confirmed on June 27 that conversations with the government are continuing, with the goal of also restoring access to Fable 5 for general use.

Conclusion

The US government’s June 27 decision to partially lift the export block on Anthropic’s Mythos 5 is a step toward stability. However, Fable 5 remains inaccessible to the public, including users and businesses in India. For India’s fintech lenders who depend on AI tools for credit and recovery, the resolution of this export dispute matters beyond US borders. Regulatory clarity on frontier AI access will be critical for the sector in the second half of 2026.

FAQs

What does the US Commerce Department's June 27 letter mean for Anthropic's Mythos 5 access?

The letter, signed by Commerce Secretary Howard Lutnick, confirms that Mythos 5 can now reach a limited group of trusted cyber defenders and infrastructure providers. The full public ban on Fable 5 remains in place. Negotiations between Anthropic and the US government are still ongoing as of June 27, 2026.

Which companies and agencies did the Trump administration allow to access Anthropic's Mythos 5 on June 27, 2026?

The US government has not named specific companies publicly. Access on June 27, 2026 is limited to a small group of cyber defenders and critical infrastructure providers. Anthropic confirmed it is working to restore their access to Mythos 5 as quickly as possible.