EU's cybersecurity agency granted access to Mythos 5 AI model, Commission says

• ENISA is now actively testing Anthropic's Mythos 5 AI model after receiving access approval
• The agency has also gained access to OpenAI's latest model, GPT-6-Astra, for evaluation
• This access enables EU cybersecurity authorities to directly assess and monitor cutting-edge AI technologies from major developers

Summary: EU Cybersecurity Agency Gains Access to Advanced AI Models

Key Development:

The European Union's cybersecurity agency ENISA has been granted access to two major AI models for testing purposes, according to the European Commission on September 10, 2026.

Companies and Models Involved:

• Anthropic: ENISA gained access to the Mythos 5 AI model
• OpenAI: ENISA also received access to GPT-6-Astra, OpenAI's latest model

Market Implications:

This development signals increased EU regulatory scrutiny of advanced AI systems, consistent with the bloc's proactive approach to AI governance. The access grants suggest both Anthropic and OpenAI are cooperating with European authorities to ensure their models meet cybersecurity and regulatory standards before broader deployment in EU markets.

The move comes as part of the EU's comprehensive AI regulatory framework, which requires thorough vetting of powerful AI systems. ENISA's testing will likely focus on security vulnerabilities, data protection compliance, and potential risks associated with these advanced models.

Related Context:

Separately, Germany's military announced it is evaluating 30 AI tools for battlefield decision-making, planning deployment from 2027, highlighting growing government interest in AI applications across defense and security sectors.

Significance for Investors:

This regulatory engagement could impact timelines for commercial deployment of these AI models in the EU market. Companies operating in the AI sector should monitor ENISA's findings, as they may establish precedents for AI model approval processes and influence regulatory requirements across the 27-member bloc. The cooperative approach from major AI firms suggests they view EU market access as strategically important despite regulatory hurdles.