Google releases Gemini 3.8 Flash, its third Flash model in six weeks

Google has released Gemini 3.8 Flash, marking the third time in six weeks that the company has rolled out a new variant of its lightweight model family. This rapid cadence suggests a strategic pivot away from the long-awaited Gemini 3.5 Pro, which has yet to materialize, toward high-frequency updates of the "workhorse" Flash models. The new release is pitched as a significant upgrade for developers, with Google claiming it performs more reasoning steps and calls tools iteratively to solve complex tasks.

For developers, the primary draw is the combination of performance and cost. Gemini 3.8 Flash is available at an introductory rate of $0.75 per million input tokens and $3.75 per million output tokens through the end of the year. While third-party analyses note that the model may use more tokens per task to maximize performance—potentially increasing total costs compared to its predecessor—it currently tops the DeepSWE leaderboard for software engineering. This places it in direct competition with larger, more expensive frontier models, including recent upgrades from Anthropic.

Alongside the standard release, Google introduced Gemini 3.8 Flash Cyber, a specialized variant tuned for vulnerability detection and mitigation. This model is not available to the general public but is launching through Google’s new Fairwind Program, a restricted initiative for governments and trusted partners like CrowdStrike and Palo Alto Networks. The Cyber variant reportedly offers a 2.6x increase in patch accuracy for the Chrome security team and can autonomously find and fix vulnerabilities using a new agent called CodeMender.

While the coding benchmarks are strong, Google’s models still lag behind market leaders in agentic computer use tests like OSWorld-2.0. Nevertheless, the aggressive release schedule and competitive pricing signal that Google is prioritizing immediate utility and developer adoption over waiting for a single, perfect frontier release.