OpenAI has initiated a new program to provide European Union cybersecurity agencies with specialized access to GPT-5.5-Cyber, a version of its artificial intelligence designed specifically for digital defense. By integrating this advanced technology into the workflows of European government security teams, the initiative aims to fortify the region’s digital infrastructure against increasingly sophisticated cyber threats. A select group of security professionals has already begun testing the model, marking a significant step in the integration of AI-driven capabilities into public sector defense strategies.

The European Commission has welcomed the collaboration, noting that direct access allows regulators to better understand and oversee the deployment of powerful AI models. This hands-on approach is expected to help authorities identify potential security vulnerabilities and develop proactive mitigation strategies. The framework for this partnership is designed to broaden the availability of high-level defensive AI tools, moving these resources beyond the walls of private technology firms and into the hands of public sector experts tasked with national and regional security.

While the partnership with OpenAI moves forward, similar efforts to secure access to other industry-leading models have faced delays. The European Commission has yet to reach an agreement for preview access to Mythos, a security-oriented model developed by Anthropic that was released publicly last month. Although negotiations between the Commission and Anthropic are reportedly in progress, they have not yet yielded a formal arrangement, and Anthropic has not provided official comments on the current status of these discussions.