OpenAI has moved proactively to strengthen ties with European regulators by offering the European Union early access to its latest cybersecurity-focused AI model, GPT-5.5-Cyber, while rival Anthropic continues to withhold preview access to its own powerful system, Mythos.
The announcement on Monday represents a notable diplomatic gesture by OpenAI amid heightened EU scrutiny of frontier AI systems, particularly those with dual-use potential in cybersecurity. Under the arrangement, European partners, including businesses, national governments, cyber defense authorities, and EU institutions such as the EU AI Office, will gain access to the specialized model.
OpenAI said it is initially rolling out GPT-5.5-Cyber in a limited preview to carefully vetted cybersecurity teams and organizations. The move comes one month after Anthropic released Mythos, a development that triggered significant concern across Europe over the potential for highly capable AI to be used in offensive cyberattacks against critical infrastructure, government systems, and private networks.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
European Commission spokesperson Thomas Regnier welcomed the step, stating at a Monday press briefing: “We welcome OpenAI’s transparency and intent to give Commission access to new model.”
He confirmed that initial exchanges had already taken place and that further technical discussions are scheduled this week.
“This will allow us to follow deployment of the model very closely, and address security concerns,” Regnier added.
In contrast, discussions with Anthropic remain at an earlier stage. Regnier noted that while the Commission has held “four or five” meetings with the company, the talks have “not yet [reached] the same stage as the solution we have on the table from OpenAI.”
George Osborne, OpenAI’s Head of OpenAI for Countries and former UK Chancellor of the Exchequer, framed the decision as part of a broader commitment to responsible AI development in Europe.
“AI labs like ours shouldn’t be the sole arbiters of cyber safety as resilience depends on trusted partners working together,” Osborne said. “The latest cyber AI capabilities should be available for Europe’s many defenders, not just the few, and we want to help make that happen.”
Through the newly launched “OpenAI EU Cyber Action Plan,” the company pledged to collaborate with European policymakers, institutions, and businesses to democratize access to defensive AI tools while aligning development with European values and security priorities.
Diverging Approaches Between AI Leaders
The contrasting responses from OpenAI and Anthropic highlight deepening differences in how the two leading American AI labs navigate European regulation. OpenAI appears to be pursuing a more collaborative and transparent approach, likely aimed at building long-term trust and avoiding harsher regulatory measures under the EU AI Act.
Anthropic, known for its strong emphasis on safety and alignment research, has taken a slower, more guarded approach to releasing advanced models in regulated markets. While this caution has earned praise from some safety advocates, it is beginning to create friction with European officials eager to assess capabilities and risks in real time.
Mythos’s release last month sparked particular alarm because of its reported strength in offensive cybersecurity tasks, including code exploitation and vulnerability discovery. European governments and critical infrastructure operators have grown increasingly wary of a scenario in which such tools fall into the wrong hands or are misused by state actors.
The development, amid pressure from Washington to cut U.S. tech companies a slack, underscores the EU’s determination to maintain oversight of powerful AI systems operating within its borders. As cyber threats from nation-states and criminal groups continue to evolve, European authorities are keen to avoid being left behind in both defensive and offensive AI capabilities.
However, the outreach serves multiple purposes for OpenAI: it helps mitigate regulatory risk, builds goodwill with key policymakers, and positions the company as a responsible partner in Europe’s digital sovereignty push. It is also expected to give the company a competitive edge in winning contracts and partnerships with European governments and enterprises.
The situation also reflects wider transatlantic tensions over technology governance. While the U.S. has traditionally favored a lighter regulatory touch, the EU continues to assert greater control through the AI Act, GDPR, and other digital regulations.
Against this backdrop, how Anthropic responds in the coming weeks could have significant implications for its standing in Europe and its broader global expansion strategy. As discussions continue, the EU will be looking not just for access, but for meaningful transparency, risk assessments, and the ability to impose safeguards if necessary.