SiliconANGLE reports that CrowdStrike unveiled autonomous red teaming during its Fal.Con 2026 keynote. The company introduced SafeMind along with its Red Tempest and Blue Solano models. Coverage states frontier AI has favored attackers and prompted CrowdStrike along with Nvidia to build SafeMind as a security model family. The models supply red and blue agents so defenders can close the resulting AI gap. The reports describe the release as the start of dedicated cybersecurity model families rather than general purpose tools.