Google Launches Gemini 3 Flash, Promises Faster Performance and Lower Costs

Google has rolled out Gemini 3 Flash, a new AI model for developers that the company claims delivers faster performance and lowers costs while retaining advanced reasoning and multimodal capabilities.

The company is making the frontier intelligence model accessible through the Gemini API via Google AI Studio, Gemini CLI, Android Studio, agentic development platform Google Antigravity, and for enterprise customers through Vertex AI. 

Gemini 3 Flash is priced at $0.50 per million input tokens and $3 per million output tokens, with additional cost reductions through context caching and batch processing.

“Today we’re introducing Gemini 3 Flash, our latest model with frontier intelligence built for speed at a fraction of the cost,” said Logan Kilpatrick, group product manager at Google DeepMind. 

According to Google, Gemini 3 Flash builds on the capabilities of Gemini 3 Pro and outperforms Gemini 2.5 Pro across several benchmarks, while operating up to three times faster. It stated that the model supports multimodal reasoning, coding, agentic workflows, and visual understanding, including code execution for tasks such as counting, zooming, and editing visual inputs.

Google said the Flash series remains its most widely used model family, processing trillions of tokens across hundreds of thousands of applications. With Gemini 3 Flash, the company aims to support large-scale production use cases that require lower latency and higher rate limits.

Early users have already integrated the model into products spanning software development, gaming, document analysis, and deepfake detection. Google also recently introduced CC, an experimental AI productivity agent developed by Google Labs, to help users manage daily tasks and organise their workday more efficiently.