Dubbed Unit 42 Continuous Frontier AI Defense, the platform pairs open-weight architectures with cybersecurity-tailored systems, including Anthropic’s Claude Mythos 5 and OpenAI’s GPT-5.6-Cyber. Palo Alto Networks has announced an enterprise-grade cybersecurity service powered by artificial intelligence (AI) models from Anthropic and OpenAI , aiming to identify and neutralise network vulnerabilities before malicious actors can exploit them. The deployment comes as cybercriminals increasingly rely on automated algorithms to locate and breach enterprise weaknesses, prompting security companies to build counter-defenses capable of matching automated hacking cycles.
The system continuously evaluates web applications, application programming interfaces (APIs), and cloud infrastructure, according to the company. Unlike traditional periodic vulnerability scans, the service operates as an always-on offensive testing engine built to adapt alongside evolving enterprise environments. It maps out complete end-to-end attack paths, allowing corporate defenders to pinpoint exposed entry points and hidden exploit vectors.
International Subscription Pricing for New Tool
The company has announced that its new tool will be available worldwide through an annual subscription model. Pricing will vary based on the specific combination of OpenAI, Anthropic, and open-source models selected by each enterprise client.