The U.S. government has officially lifted worldwide export restrictions on Anthropic’s Claude Fable 5 and Mythos 5 AI models, ending an 18-day suspension that temporarily blocked access for users outside the United States.
Anthropic confirmed the decision on Tuesday, saying it will begin restoring access to both models in phases across its platform and cloud partners. The move follows weeks of discussions with the U.S. Commerce Department over the security risks posed by the company’s latest frontier AI systems.
Why the Restrictions Were Introduced
The restrictions were first imposed on June 12 after U.S. officials raised national security concerns about Claude Fable 5’s advanced cybersecurity capabilities.
According to Reuters, researchers demonstrated a jailbreak that allowed the model to identify software vulnerabilities and generate exploit code. Because Anthropic could not reliably verify the nationality of every user in real time, the Commerce Department ordered the company to suspend global access to Fable 5 and Mythos 5 rather than risk the technology reaching restricted foreign users.
The decision reflected a broader shift in Washington’s approach to frontier AI. Instead of treating advanced models as ordinary software, regulators began viewing them as technologies with potential national security implications.
Anthropic Worked With Regulators Instead of Fighting the Decision
Following the suspension, Anthropic said it believed the order was based on a misunderstanding and apologized to customers for the disruption.
Rather than challenge the restrictions publicly, the company worked closely with the Commerce Department to strengthen its safety systems. Anthropic introduced new protections that can reportedly block the jailbreak technique in about 99% of tested cases. Suspicious prompts are now redirected to a safer model, while additional monitoring teams will watch for misuse around the clock.
Anthropic also committed to sharing information about malicious activity with the U.S. government, providing pre-release access for security evaluations, and helping develop industry-wide AI safety standards through Project Glassing alongside Amazon, Microsoft, and Google.
OpenAI Also Felt the Impact of the New Oversight
Anthropic was not the only AI company affected by Washington’s tougher oversight.
Reuters reported that OpenAI also delayed the full public launch of GPT-5.6 at the U.S. government’s request, limiting access to a smaller group of vetted partners while additional safety reviews were carried out. OpenAI CEO Sam Altman said safety testing is valuable but questioned whether governments should decide which customers receive access to frontier AI models.
The developments suggest the new oversight framework applies across the AI industry rather than targeting a single company.
How the Restrictions Were Lifted
The Commerce Department removed the export controls after reviewing Anthropic’s updated safeguards and reaching an agreement with the company.
In a post announcing the decision, U.S. Commerce Secretary Howard Lutnick said officials had worked closely with Anthropic over the previous two weeks to ensure the models aligned with national security objectives while supporting America’s leadership in AI. Anthropic later confirmed that it had received official notice that the export controls had been lifted and that global access would begin returning on July 1.
What the Decision Means
The removal of the restrictions allows Anthropic to resume the global rollout of Claude Fable 5 and Mythos 5, giving developers, researchers, and businesses outside the United States access to the company’s latest AI models once again.
The decision is also significant for the broader AI industry. It shows that the U.S. government is not seeking to permanently block frontier AI models from international markets. Instead, regulators appear willing to approve global deployment when companies demonstrate stronger security protections and closer cooperation on AI safety.
As competition between Anthropic, OpenAI, Google, and other AI developers accelerates, the episode may become a blueprint for how future frontier AI systems are reviewed before they are released worldwide.
READ ALSO: OKX Unveils AI Marketplace Where Autonomous Agents Can Hire and Pay Each Other
Google makes Gemini’s Personalized AI Image Generation free for U.S. Users
MTN Confirms Jerry Soko as Substantive CEO of Eswatini Unit
Idris Elba and Google Partner to Empower 100,000 African Creators with AI Tool
