OpenAI announced Monday that it will grant the European Union access to its new cybersecurity AI model, GPT-5.5-Cyber, while rival Anthropic has yet to offer the bloc similar access to its own cybersecurity model, Mythos. European Commission spokesperson Thomas Regnier said the Commission welcomed OpenAI's engagement and noted that despite having met with Anthropic on four or five occasions, no conversations about obtaining access to Anthropic's AI models have taken place, describing the Anthropic talks as being at a “different stage.” Access to GPT-5.5-Cyber will be extended to European businesses, governments, cybersecurity authorities, and EU bodies including the EU AI office through a limited preview aimed at vetted cybersecurity teams. The move comes about a month after Anthropic announced Project Glasswing, which restricted Mythos access to just 12 launch partners (including AWS, Apple, Cisco, CrowdStrike, Google, Microsoft, and Nvidia) plus more than 40 additional organizations, all of whom agreed to use the model exclusively for defensive security work.
