Google is introducing a cheaper alternative to big AI security brands like Mythos

Google is launching an AI security model dedicated to quickly finding and identifying security threats. In Tuesday’s blog postGoogle describes Gemini 3.5 Flash Cyber ​​as a “low-cost and highly capable alternative” to large, expensive AI systems, such as those offered by Anthropic’s Mythos.

The new version is built on Gemini 3.5 Flash and will be available first to governments and trusted partners through CodeMender, Google’s security assistant. As stated by Google, CodeMender can call 3.5 Flash Cyber ​​”multiple times at high speed and low cost,” allowing AI agents to analyze multiple code paths and find problems.

Google says 3.5 Flash Cyber ​​performed “competitively” against “the largest brands” on the CyberGym AI cybersecurity benchmark after being called up to five times. It also identified “55 confirmed vulnerabilities” in the V8 JavaScript Engine, compared to 47 found with Gemini 3.5 Flash and 36 with Opus 4.6. Google adds that 3.5 Flash Cyber ​​found 10 issues that no other model found, noting that the model continued to find new systems and problems after several requests.