Google Launches Gemini 3.5 Flash Cyber to Rival Mythos
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Google launched Gemini 3.5 Flash Cyber, describing it in a blog post as a "cost-efficient and highly capable alternative" to larger, more expensive AI security systems like Anthropic's Mythos.
- Gemini 3.5 Flash Cyber will roll out first to governments and trusted partners through CodeMender, Google's security-focused coding agent, which can call the model "multiple times at high speed and low cost" to scan more code paths.
- 3.5 Flash Cyber detected 55 unique confirmed issues in Google's V8 JavaScript Engine, outperforming Gemini 3.5 Flash (47) and Claude Opus 4.6 (36), and flagged 10 vulnerabilities no other model discovered, per Google's CyberGym benchmark testing.
- Anthropic's Mythos 5, released under the Project Glasswing initiative, costs twice as much as Claude Opus 4.8 to run, and Microsoft — which adopted Mythos for security checks — just logged its biggest Patch Tuesday after using AI to surface vulnerabilities.
- China's Z.ai also claims its model can compete with Mythos, signaling that the race for affordable AI cybersecurity tools now extends well beyond the U.S. frontier-model labs.
- Google separately unveiled an upgraded Gemini 3.6 Flash with coding and multimodal gains, plus 3.5 Flash-Lite as the "most cost-effective" entry in the 3.5 family.
Why it matters: Google's 3.5 Flash Cyber detected 55 unique V8 JavaScript Engine vulnerabilities — beating Claude Opus 4.6's 36 — while undercutting Anthropic's Mythos 5, which runs at 2x Claude Opus 4.8 pricing. With Microsoft already locked into Mythos for security checks and posting its biggest Patch Tuesday this month, Google's low-cost model lands as a direct competitive alternative for governments and enterprise security teams weighing Mythos adoption.
