Introducing Gemini 3.8 Flash and 3.8 Flash Cyber

Building on the momentum of 3.7 Flash released three weeks ago, Google DeepMind today introduced Gemini 3.8, marking its third Flash release in six weeks. The new model is positioned as the company's best reasoning and coding model yet, offering the same speed and low cost as 3.7 Flash. It comes in two variants: Gemini 3.8 Flash, an upgraded general-purpose 'workhorse' model, and Gemini 3.8 Flash Cyber, a specialized cybersecurity model with frontier-level vulnerability detection and automated patching capabilities, available to trusted defenders through the new Fairwind Program.

Both variants share the same foundational intelligence and are accelerated by long-running agentic loops that recursively evaluate and refine the underlying models. Significant coding and reasoning gains were driven by innovations including rigorous training in the demanding domain of cybersecurity. On DeepSWE v1.1, a long-horizon software engineering benchmark, 3.8 Flash outperforms most larger frontier models at autonomously solving complex engineering problems end to end, but at a fraction of the cost. It also shows strong results in specialized professional domains, outperforming 3.7 Flash and other frontier models on benchmarks like Vals Finance Agent V2 and Harvey's Legal Agent Benchmark, and achieves a 54.9% score on HLE-Verified for multi-step reasoning across STEM, humanities, and professional fields.

The performance gains stem from a core design choice: 3.8 Flash works harder, executing extra reasoning steps and iteratively calling tools on complex tasks, which may use more tokens at higher effort levels. Developers with compute-efficiency constraints can use lower effort levels to minimize token overhead or continue using Gemini 3.7 Flash, which remains fully supported. Pricing for 3.8 Flash matches 3.7 Flash at $0.75 per million input tokens and $3.75 per million output tokens. Gemini 3.8 Flash Cyber is available to a set of trusted defenders through the Fairwind Program.