Google announces Sec-Gemini v1, a new experimental cybersecurity model

The document announces Sec-Gemini v1, an experimental AI model that integrates Gemini with Google Threat Intelligence, OSV, and Mandiant data to enhance SecOps tasks such as incident root cause analysis, threat analysis, and vulnerability understanding; it reports superior results on CTI-MCQ and CTI-RCM benchmarks and invites select organizations to request early research access.