Google drops Gemini 3.8 Flash and a specialized Cyber model

Google is shipping models faster than most engineering teams update their dependencies.

Just three weeks after launching 3.7 Flash, Google DeepMind announced Gemini 3.8 Flash alongside a security-focused 3.8 Flash Cyber. It marks Google's third Flash release in six weeks. The core 3.8 Flash model keeps the price at $0.75 per million input tokens while boosting multi-step reasoning and software engineering performance.

Why it matters: The standard 3.8 Flash punches well above its weight by simply working harder. On complex tasks, it executes extra reasoning steps and calls tools iteratively to beat larger models on long-horizon benchmarks like DeepSWE v1.1. For security teams, the Cyber variant offers frontier-level vulnerability discovery and automated patching at a fraction of typical frontier costs—helping Chrome's security team generate 2.6 times more correct patches than larger commercial alternatives.

Try it / know this: You can start building with 3.8 Flash now in the Gemini API or Google Antigravity. If you need lower token overhead, you can reduce the model's effort level or stick with 3.7 Flash, which remains fully supported. The Cyber model is available to vetted defenders through Google's new Fairwind Program.

At this release pace, Gemini 3.9 might drop before you finish reading this sentence.