Sec-Gemini Docs
Framework-Agnostic Web Component
Frequently Asked Questions
What is Sec-Gemini?
Sec-Gemini v1 is a new experimental AI model from Google, focused on advancing cybersecurity AI frontiers. It aims to help shift the balance back to defenders by force multiplying cybersecurity professionals. It combines Gemini’s advanced capabilities with near real-time cybersecurity knowledge and tooling to achieve superior performance on key cybersecurity workflows.
Who can access Sec-Gemini v1?
Google is making Sec-Gemini v1 freely available to selected organizations, institutions, professionals, and NGOs for research purposes. This is part of Google’s commitment to strong collaboration across the cybersecurity community to push AI cybersecurity frontiers.
What is the purpose of making Sec-Gemini v1 freely available for research?
The purpose is to foster strong collaboration across the cybersecurity community and push AI cybersecurity frontiers to decisively tilt the balance in favor of defenders. By making it available for research, Google aims to gather feedback and explore new use cases for AI in cybersecurity.
How do I get an API key?
To get API keys for Sec-Gemini, visit the key management UI. You will need to already have access to Sec-Gemini and log in with your Google account to access this page.
How do I access the UI?
You can access the Sec-Gemini UI at secgemini.google. You will need to log in with your Google account that has access to Sec-Gemini.
What are the key capabilities of Sec-Gemini v1?
Sec-Gemini v1 is designed to enhance cybersecurity workflows by offering state-of-the-art reasoning capabilities and extensive current cybersecurity knowledge. Its key capabilities include incident root cause analysis, threat analysis, and vulnerability impact understanding. It integrates with Google Threat Intelligence (GTI), OSV (Open Source Vulnerabilities), and other key data sources to provide near real-time insights.
How does Sec-Gemini v1 perform compared to other models?
Sec-Gemini v1 outperforms other models on key cybersecurity benchmarks. It scores at least 11% higher on CTI-MCQ (Cyber Threat Intelligence – Multiple Choice Questions), a leading threat intelligence benchmark, and at least 10.5% higher on the CTI-Root Cause Mapping benchmark.