In a nutshell: Anthropic is permitted to release its Claude 5 model to selected US cyber defenders following security reviews, while weaker variants remain subject to export restrictions.
The US government has allowed Anthropic to release its AI model Claude 5 to selected cyber defenders and infrastructure providers. The model has been subject to a complete export ban since June due to security concerns.
Commerce Secretary Howard Lutnick informed Anthropic in writing that the required security safeguards have been implemented and partial release of Claude 5 is now approved. This grants selected trusted companies access to the model once again. The weaker model Claude 3.5 Sonnet, however, remains blocked. According to Anthropic, the company is conducting ongoing discussions with the government regarding possible release of this model as well.
In early June, Washington imposed an export ban after security experts warned that the model could be used by hackers to identify and exploit IT vulnerabilities unusually quickly. To comply with this order, Anthropic had completely shut down access to both models – even for its own foreign employees.
The partial release demonstrates the pragmatic approach of the US government, which is attempting to balance national security concerns against the economic interests of American AI companies in competition with China. At the same time, the approach illustrates how fragmented the regulatory framework for AI systems in the US currently is. Additional tensions arose when the government classified Anthropic as a supply chain risk due to disagreements over military applications – Anthropic sued and, according to its own statements, achieved an initial partial court victory.
Source: www.it-daily.net · Published 29 June 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.2.