Introducing Gemini 3.5 Flash Cyber

Google has launched Gemini 3.5 Flash Cyber, a specialized AI model built on Gemini 3.5 Flash, to enhance cybersecurity defense by quickly identifying, validating, and patching software vulnerabilities. This lightweight model demonstrates superior performance and cost-efficiency compared to larger cybersecurity models, proving effective in benchmarks and real-world applications within Google’s own codebases. Initially available to governments and trusted partners via CodeMender, it aims to empower defenders with a scalable and affordable solution against evolving cyber threats.

• Gemini 3.5 Flash Cyber is a new, lightweight cybersecurity AI model from Google, fine-tuned for rapid vulnerability detection, validation, and patching.
• It is built on the efficient Gemini 3.5 Flash architecture, offering a cost-effective alternative to larger, more expensive cybersecurity models.
• The model has shown competitive and superior performance against larger models in various benchmarks, including CyberGym and Google’s internal Big Sleep evaluation.
• Gemini 3.5 Flash Cyber is already being used within Google’s internal codebases, including Chrome, Android, and Cloud, to find and fix vulnerabilities.
• Access to Gemini 3.5 Flash Cyber will initially be limited to governments and trusted partners through a pilot program via CodeMender, with broader expansion planned.
• The model’s speed and affordability allow for integration into frequent scans, time-sensitive processes, and large-scale commit scanning pipelines.
  Continue reading https://deepmind.google/blog/introducing-gemini-3-5-flash-cyber/