EU cyber agency finally testing Anthropic's Mythos 5 | Euractiv

But US AI lab's latest Mythos 5.1 cyber model is still being withheld

The EU’s cybersecurity agency ENISA is finally testing Anthropic’s cyber-focused AI model, Mythos 5, months after the bloc first sought to secure access.

“Following our constructive engagement with Anthropic, we can confirm that the EU’s cybersecurity agency ENISA has been granted access to Mythos 5 and is testing it now,” a Commission spokesperson told Euractiv in a statement.

The development, first reported by Bloomberg and confirmed to Euractiv by the Commission, comes days after Anthropic released an updated version of the cyber model: Mythos 5.1.

In that case, Anthropic has indicated that the White House has not allowed it to share the model outside the US – meaning that the EU remains locked out of accessing Mythos 5.1.