EU and Anthropic discussions continue on Mythos access

Anthropic and the European Union continue to engage in discussions regarding access to the company's advanced artificial intelligence model, Mythos. The European Commission has sought access to the model due to cybersecurity risks associated with its capabilities. The EU had to obtain permission from the U.S. administration to access the technology, as the American government has reluctance to share powerful models with non-U.S. governments.

Anthropic initially launched Mythos in April as part of its cybersecurity initiative, Project Glasswing, with the model demonstrating an ability to identify software vulnerabilities. The release of Mythos raised alarms among governments, banks, and tech firms due to its potential for misuse by malicious actors.

The EU has already been granted access to OpenAI's GPT-5.5-Cyber model, but discussions with Anthropic remain ongoing. While the EU has had several meetings with Anthropic, the official terms of the agreement have not yet been disclosed. The European Commission emphasized the importance of understanding the risks posed by advanced AI models and expressed a commitment to continued collaboration with international partners.