Palo Alto Networks Launches Unit 42 Continuous Frontier AI Defense

Palo Alto Networks launched Unit 42 Continuous Frontier AI Defense, an offensive security service for enterprises, the company said in a press release. The service uses Claude Mythos 5 from Anthropic, GPT-5.6-Cyber from OpenAI, and other security focused models to identify and validate exposures continuously.

The service tests web applications, APIs, cloud infrastructure, source code repositories and network assets. It provides prioritized remediation guidance, including code fixes and virtual patch recommendations.

Palo Alto Networks said it developed the service through six months of internal testing and more than 100 Unit 42 customer engagements, supported by $17 million in research and development spending. The company said every assessed customer had exposures, with 37% classified as high or critical severity.

The service is available worldwide through annual subscriptions. Customers can select options based on models from OpenAI, Anthropic and open source providers.