Palo Alto Networks partners with OpenAI, Anthropic to deliver Unit 42 Continuous Frontier AI Defense service

Palo Alto Networks launched Unit 42 Continuous Frontier AI Defense, integrating Anthropic’s Claude Mythos 5 and OpenAI’s GPT-5.6-Cyber models.

The service targets continuous offensive testing to find, validate, and remediate exposures before attackers weaponize them, using a proprietary multi-model harness.

It will be available worldwide via annual subscription, with options varying by the OpenAI, Anthropic, or open-source models selected.

Palo Alto Networks cited $17 million of R&D investment, with customer assessments finding exposures in 100% of engagements.