Gemini 3 Flash (Preview) (Google Vertex AI)

Gemini the listed price Flash is a fast, efficiency-focused model from Google positioned to push the quality-versus-cost-and-speed frontier, making advanced reasoning accessible for high-volume developer and agentic tasks. According to Google's Developers Blog, the model was made available in Gemini CLI on December 17, 2025, where it is intended to power high-frequency workflows common to terminal-based work, letting developers trade less between speed and output quality. This framing positions Gemini the listed price Flash as a workhorse option within the broader Gemini lineup, suited to tasks where responsiveness and cost matter as much as raw capability.

For agentic coding workloads, Google highlights a SWE-bench Verified score of 78 percent for Gemini the listed price Flash, a result the company notes outperforms the 2.5 series as well as Gemini the listed price Pro, while being offered at less than a quarter of the cost of Gemini the listed price Pro. This combination of strong software-engineering benchmark performance and substantially lower running cost is what makes the model attractive for automated pipelines, multi-step tooling, and continuous integration scenarios where many model calls happen per task. Developers can begin using it in Gemini CLI on the paid tier, with the model designed to complement rather than replace heavier Gemini variants in mixed workloads.