The US Department of Commerce has lifted the export controls that had applied to Anthropic’s Fable 5 and Mythos 5 artificial intelligence models. The decision ends a standoff between the company and the Trump administration.
The move was announced by Commerce Secretary Howard Lutnick. According to the statement, the Bureau of Industry and Security (BIS) withdrew the restrictions that applied to the two models.
Anthropic had launched Fable 5 and Mythos 5 on June 9, 2026. A few days later, the government imposed the suspension, which then remained in place for about 19 days.
The Mid-June Suspension
Anthropic had disabled access to Fable 5 and Mythos 5 on June 12 to comply with a government directive. The directive cited national security authorities as the basis for the order.
The company was told to suspend all access by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The measure interrupted use of the models until July 1.
The order has been described as one of the most far-reaching restrictions ever imposed by a government on an AI model. During the suspension, many users and developers fell back to earlier models such as Opus 4.8.
Restoring the Models
On June 26, Secretary Lutnick had partially restored Mythos 5 for organizations involved in defending critical infrastructure. The full removal of the controls came on June 30.
After the controls were lifted, Fable 5 became available again to users worldwide on the Claude platform, on Claude.ai and on Claude Code starting July 1, 2026.
Access to Mythos 5 was restored for some US organizations. The full return of the two models therefore happened in stages.
The Origin of the Case and New Safeguards
The restriction originated from a jailbreak technique identified by Amazon researchers. The method bypassed the model’s safety classifiers and, in one case, led it to produce code showing how to exploit a vulnerability.
According to Anthropic’s later testing, every other model it examined could produce the same output, including Claude Haiku 4.5, Sonnet 4.6, Opus 4.6 through 4.8, GPT-5.4, GPT-5.5 and Kimi K2.7.
For the restoration, Anthropic trained a new safety classifier, which the company says blocks the technique in more than 99% of attempts. The Commerce Department’s Center for AI Standards and Innovation evaluated the new safeguards before lifting the controls.
Anthropic also made four commitments to the government: early access to future frontier models, rapid reporting of jailbreaks and misuse, sharing of new safeguards, and developing, with Amazon, Microsoft and Google, a system to assess how dangerous a jailbreak is. (Source: https://www.forbes.com/sites/siladityaray/2026/07/01/trump-administration-lifts-export-controls-on-anthropics-mythos-5-and-fable-5-ai-models/)
Original article: cnbc.com
