The essentials: Anthropic’s Claude-3.5-Sonnet model is cleared for distribution to over 100 Trusted Partners, while Claude-3.5-Opus remains blocked and the government develops a standardized assessment framework for future security disputes.
On Friday, the Trump Administration partially lifted export restrictions on Anthropic’s Claude-3.5-Sonnet model, de-escalating a confrontation from nearly two weeks earlier. The model can now be distributed again to a group of over 100 trusted companies and federal agencies, while the advanced Claude-3.5-Opus model remains blocked.
The Trump Administration justified the export restrictions with cybersecurity concerns, fearing that the AI software could be misused for cyber attacks. Anthropic disputed these concerns and announced it would restore Claude-3.5-Sonnet access for approved partners as quickly as possible. At the same time, the company is working with the government to expand access to Claude-3.5-Sonnet and to secure a release of Claude-3.5-Opus for general use.
Bilateral discussions between Anthropic and the Administration are also focused on developing a standardized framework to assess future cases of alleged security violations. Commerce Secretary Howard Lutnick commended Anthropic in his Friday letter for “substantial progress” in addressing risks associated with the affected models. According to an informed source, the discussions will continue over the weekend.
In parallel, White House pressure is also affecting OpenAI. The company announced on Friday that, at the Administration’s request, it is undertaking only a limited release of three versions of its new GPT-5.6 model. However, OpenAI stated that it does not believe this government access process should become the long-term standard, as it keeps the best tools away from users, developers, businesses, and cyber defense forces.
This series of measures reveals tensions in technology regulation: while Trump signed an Executive Order in early June rejecting mandatory federal AI controls and instead favoring voluntary security reviews by the technology industry, the government intervention in model releases now signals that the Administration is intervening more actively in the deployment of U.S. AI products than previously announced.
Source: www.politico.com · Published 27 June 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 of the EU AI Act. Paraphrase and classification through Lumi News Pipeline v1.7.1.