In the Wake of Anthropic's Mythos, OpenAI Has a New Cybersecurity Model—and Strategy

OpenAI says its safeguards “sufficiently reduce cyber risk" for now, while GPT-5.4-Cyber is a new cybersecurity-focused model.