Google DeepMind has announced Gemini 3.5 Flash Cyber, a new model designed for cybersecurity work. The announcement describes it as lightweight and says it is intended to find and patch vulnerabilities.
The brief release does not include details on the model's architecture, training data, or availability. It does signal that Google is applying its Gemini line to specialized security tasks, where automated vulnerability detection and remediation are of growing interest.
For security teams, a lightweight model could be easier to integrate into existing tooling than larger general-purpose systems. More specifics will depend on future documentation from Google DeepMind.