OpenAI launches GPT-5.5-Cyber model to match Anthropic's cybersecurity offering

OpenAI released GPT-5.5-Cyber, a model that can detect vulnerabilities in most software, in response to Anthropic's Mythos. Both models will be made available only to select organizations to help them patch security problems.

Access is limited to a small group of institutions; almost all organizations with access to Mythos have been US-based, leaving Europe concerned about lagging behind as open-source cyber-capable models become available to hostile actors.

The EU's strict tech regulation and consideration of banning US tech giants from handling government data may discourage those firms from strengthening European cybersecurity capabilities.