Story perspectives

OpenAI Gives EU Limited Access to GPT-5.5-Cyber

5/12/2026

Summary

OpenAI Grants EU

OpenAI will give the European Union access to its GPT-5.5-Cyber model, a cybersecurity-focused version of its latest AI system.
The model is released in a limited preview for vetted cybersecurity teams.
EU Commission spokesperson Thomas Regnier said the EU welcomed OpenAI’s transparency and will continue discussions on access this week.
The EU hopes the access will let officials monitor deployment closely and address security concerns.

Summary

Lawmakers Push Mythos

A coalition of 30 European lawmakers urged ENISA to acquire Anthropic’s Mythos model for security testing.
Four member states, including Spain, asked the European Commission to coordinate information sharing on Mythos.
Former Chancellor George Osborne warned that letting a single frontier-AI firm dominate cyber-defence could concentrate power and reduce oversight.
Critics say the digital-rules delay and a proposed industry exemption could weaken the EU’s original cybersecurity legislation.

History

Historical Context

After the EU enacted the 2016 NIS Directive, obliging 27 member states and over 500,000 critical operators to adopt risk-management, the 2022 NIS 2 revision expanded scope and allocated €1.5 billion for coordination. This regulatory evolution echoes the global reach of the U.S. NIST Cybersecurity Framework, reinforcing a continent-wide push for standardized threat detection and resilience.

1 of 3

OpenAI Grants EU

OpenAI will give the European Union access to its GPT-5.5-Cyber model, a cybersecurity-focused version of its latest AI system.
The model is released in a limited preview for vetted cybersecurity teams.
EU Commission spokesperson Thomas Regnier said the EU welcomed OpenAI’s transparency and will continue discussions on access this week.
The EU hopes the access will let officials monitor deployment closely and address security concerns.

1 / 3