Google DeepMind·2026-07-17 23:00· 2026-07-17

Introducing Gemini 3.5 Flash Cyber

AI summary

Google DeepMind released Gemini 3.5 Flash Cyber, a lightweight cybersecurity model fine-tuned from 3.5 Flash for rapid vulnerability discovery, validation and patching. Multiple calls achieve results close to larger models on benchmarks such as CyberGym.

Selection record

Threshold 60Official, first-handFirst 71Second 71

AdmittedSum of both 142 ≥ twice the threshold 120

Source tier: Official, first-hand; this tier's threshold is 60
Pre-filter: passed：发布Gemini 3.5 Flash Cyber安全AI模型
Why it was chosen: Benchmark comparisons and repairs to real codebases explain the lightweight security model’s role and deployment limitations.

A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.

Source: Google DeepMind · deepmind.google

#Model releases #Safety and alignment #Agents #Google