Sec-Gemini v1 – Google Released a New AI Model for Cybersecurity

Google announced Sec-Gemini v1, an experimental AI model built on Gemini to enhance cybersecurity operations by integrating near real-time threat intelligence from Google Threat Intelligence, OSV, and Mandiant. The model reportedly outperforms peers on CTI benchmarks, demonstrates practical value with tasks like identifying and contextualizing the “Salt Typhoon” threat actor and related vulnerabilities, and will be offered via limited free research access to selected organizations to foster industry collaboration.