Google DeepMind Launches Gemini 3.5 Flash Cyber AI for Automated Vulnerability Discovery and Patching

Google’s DeepMind has announced the release of Gemini 3.5 Flash Cyber, a specialized AI model designed to discover, validate, and patch software vulnerabilities efficiently. The model is built atop Gemini 3.5 Flash and will be exclusively available to governments and trusted partners via CodeMender, an AI-powered agent for vulnerability discovery and patching unveiled in October 2025.

According to DeepMind, the lightweight model is cost-efficient and highly capable, allowing CodeMender to call upon it multiple times at high speed and low cost to scan more code paths and find vulnerabilities. In evaluations, 3.5 Flash Cyber outperformed Gemini 3.5 Flash, 3.6 Flash, and Anthropic Claude Opus 4.6 in unearthing new vulnerabilities in codebases. Stress-testing on complex projects like Google Chrome and Apple Safari showed it significantly surpassed these models.

When tested on the V8 JavaScript Engine, Gemini 3.5 Flash Cyber found 55 unique confirmed issues, compared to 47 by Gemini 3.5 Flash and 36 by Opus 4.6, including 10 issues no other model caught. The model also produced a 100% reliable remote-code execution exploit that bypassed standard mitigation techniques like ASLR and W^X.

Google is also bringing CodeMender’s foundational capabilities to customers through the Gemini Enterprise Agent Platform. The release of 3.5 Flash Cyber comes alongside Gemini 3.6 Flash and 3.5 Flash-Lite, optimized for improved coding and low-latency tasks.