On Monday, OpenAI announced it will grant access to its latest cybersecurity model, GPT-5.5-Cyber, to EU institutions and partners. This variant of its newest AI model will be made available in limited preview form to vetted cybersecurity teams. The collaboration encompasses EU enterprises, government agencies, cybersecurity authorities, and the European Union's Office on Artificial Intelligence. This decision comes one month after Anthropic released its cybersecurity model, Mythos. That launch had sparked external concerns about potential AI-driven cyberattacks targeting critical software. EU Commission spokesperson Thomas Regnier stated at Monday's regular press briefing: "We welcome OpenAI's transparency as well as its intent to provide the Commission with access to this new model." He confirmed that discussions between OpenAI and the EU have already taken place, with further talks scheduled later this week regarding arrangements for model access. Regnier added: "This will enable us to closely monitor the deployment of the model and address security-related concerns." However, despite Mythos being released a month ago, Anthropic has yet to offer the EU review privileges for a preview period. According to Regnier, while the EU is engaged in discussions with Anthropic, progress remains at a different stage than previously achieved with OpenAI. He noted that although four to five meetings have occurred between both parties, they have not yet reached an agreement comparable to what was finalized with OpenAI. In a statement, George Osborne, Head of Public Policy at OpenAI and former UK Chancellor of the Exchequer, remarked: "AI labs should not remain the sole arbiters of cybersecurity, since network resilience depends on collaboration among trusted partners. Emerging capabilities in network-focused AI must serve Europe’s defenders rather than just a select few." Osbourne also unveiled the "OpenAI EU Cybersecurity Action Plan," pledging cooperation with European policymakers, institutions, and businesses in the field of cybersecurity. The initiative aims to equip trusted entities with defensive tools to strengthen shared security, support public safety, and reflect European priorities.