OpenAI and Anthropic Cybersecurity

GPT-5.4-Cyber is designed to assist companies in finding and fixing software vulnerabilities before they can be exploited.

OpenAI stated that the model will initially be shared only with a limited group of trusted corporate partners to mitigate safety risks.

The launch follows concerns from global regulators and bank governors about the destructive potential of advanced AI agents.

Both companies are now positioning their technology as essential tools for national security, even as they face lawsuits and public backlash over the broader impacts of their AI development.