AI giants now competing on cybersecurity-specific models

Anthropic released Mythos, a cybersecurity-focused model, just before OpenAI added GPT-5.6-Cyber to its Daybreak service on August 10, 2026. The near-simultaneous launches signal that leading AI providers now treat security as a distinct category worth purpose-built models, not merely a use case for general-purpose systems.