CrowdStrike Launches Frontier Models for Cybersecurity, Created with NVIDIA
CrowdStrike SafeMind, the first agentic system for defenders, built with NVIDIA Nemotron, protects more and costs less than generic frontier models
AUSTIN, Texas & LAS VEGAS, September 01, 2026--Fal.Con 2026 – CrowdStrike (NASDAQ: CRWD) today introduced CrowdStrike SafeMind, a family of purpose-built security models and harnesses from the CrowdStrike Cyber Superintelligence Lab. The SafeMind agentic system will operate natively in the CrowdStrike Falcon® platform. Trusted access for standalone models and harnesses will be part of the Project QuiltWorks program.
SafeMind – Delivering Cybersecurity's First Agentic System Built for Defenders
Unlike frontier AI models, SafeMind runs as one system designed to deliver AI safety: an offensive model that finds the attack path, a defensive model that closes it, and the harnesses that operate both in the same loop.
SafeMind model training data comes from CrowdStrike Falcon sensor telemetry, the world's largest pureplay cyber dataset and edge install base. Model training also includes CrowdStrike's threat intelligence, Falcon Complete MDR event annotations, and fifteen years of incident response fieldwork, where human responders stopped breaches on the front lines. SafeMind launches with two distinct models, delivering unique security competencies:
Red Tempest: Offensive red team model release, built for advanced attack scenarios, emulating AI adversaries.
Blue Solano: Defensive blue team model release, built for protecting enterprise assets by deploying battle-tested measures that defenders use in real-life.
CrowdStrike builds the models using NVIDIA Nemotron open models in collaboration with NVIDIA, its AI design partner. The program also includes CoreWeave's AI Cloud for training and inference.
SafeMind's harnesses operationalize both the red and blue CrowdStrike models in a closed-loop system that continuously pits the models against each other to improve. The harnesses also work with frontier and open-source models, maximizing user choice and model preference while maintaining cost control. Frontier labs can tell a defender a risk exists. CrowdStrike goes beyond with the harnesses that can autonomously act on risk.
"The future of cybersecurity won't be defined by AI that simply identifies threats, it will be defined by AI that defeats them," said George Kurtz, CEO and founder of CrowdStrike. "SafeMind brings offensive and defensive models together in a system trained on CrowdStrike's unique cyber data. It finds weaknesses, strengthens protection, and gets smarter with every cycle, advancing our mission to stop breaches at machine speed."