Compliance regulation in the field of artificial intelligence is seeing substantial progress. Renowned AI company Anthropic has recently officially granted access to its Mythos AI model to the European Union's cybersecurity authority, marking a critical step for this cutting-edge large model in entering the European market and addressing local regulations.

The background of this open access stems from in-depth communication and negotiation between the two parties. Thomas Regnier, a spokesperson for the European Commission, publicly stated that after constructive discussions with Anthropic, the European Union Agency for Cybersecurity (ENISA) has been formally authorized to access the Mythos5 model and is currently conducting related tests. This development reflects the deepening of safety assessments of cutting-edge AI technologies by European regulatory authorities.

However, this open access is not for the full version. According to details revealed by Thomas Regnier, the spokesperson for the European Commission, although ENISA has access to the Mythos5 model, it still cannot access the latest version of the series, Mythos5.1. This detail indicates that there is a delicate division between the technical safety assessment and the iteration rhythm of large AI models during rapid development, and also reflects the ongoing scrutiny by regulatory authorities on the transparency and security of the latest and most advanced versions of large models.