OpenAI expanded access to GPT-5.5-Cyber, a defensive variant tuned for penetration testing, vulnerability research, and malware analysis, to dozens of European organizations under its Trusted Access for Cyber programme. Named partners include Deutsche Telekom, BBVA, Telefonica, and Sophos. Anthropic has notably not offered equivalent access to Mythos for EU review, creating a competitive gap in the AI cybersecurity market.