Google has announced the release of Gemini 3.8 Flash alongside a specialized, gated variant designated Gemini 3.8 Flash Cyber, tailored for vetted security teams and critical infrastructure operators.
Google CEO Sundar Pichai highlighted that the model delivers significant capability gains over its predecessor, Gemini 3.7 Flash, particularly in software engineering, multi-step reasoning, and agentic workflows. Testing on the DeepSWE v1.1 benchmark indicates that the model outperforms many larger frontier models in solving complex engineering tasks end to end while maintaining a lower operational cost.
According to Google senior director of product management Tulsee Doshi and Gemini Security Lead Raluca Ada Popa, performance increases stem from deeper operational execution. “3.8 Flash works harder,” Doshi and Popa noted, pointing to increased reasoning steps and repeated tool calls prior to finalizing responses. Developers requiring speed over depth retain the option to adjust the model effort settings or continue utilizing Gemini 3.7 Flash.
Deployment of Gemini 3.8 Flash Cyber is managed via a newly established framework called Fairwind. Access is strictly limited to verified government authorities, critical infrastructure operators, and software maintainers conducting vulnerability assessments on extensive codebases. “We have invested in vulnerability fixing from the start,” Doshi and Popa stated, emphasizing defensive patching capabilities over offensive tooling.
Metrics provided by Chrome Security indicate that Gemini 3.8 Flash Cyber yields 2.6 times more accurate patches than comparable commercial models. Furthermore, Google’s Cloud Vulnerability Research team reported identifying a critical foundational vulnerability in under two hours utilizing the platform. Standard iterations of Gemini 3.8 Flash ship with built-in safeguards restricting CBRN misuse and cyber-offense functions, whereas the Cyber edition incorporates more permissive cybersecurity parameters requiring strict access controls.
Gemini 3.8 Flash is commercially available starting at introductory rates of $0.75 per million input tokens and $3.75 per million output tokens. Integration options span the Gemini API within Google AI Studio, Google Antigravity, Android Studio, and Stitch, with enterprise deployment supported through Gemini Enterprise and consumer availability via Google AI Pro and Ultra subscriptions.