OpenAI opens its cybersecurity model to thousands of defenders in race with Anthropic’s Mythos
OpenAI launches GPT-5.4-Cyber with binary reverse engineering for verified defenders, scaling access to thousands as it competes with Anthropic's restricted Mythos model.