Gemini 2.5 rollout: Google makes Flash and Pro stable, previews Flash-Lite

Google has expanded its Gemini 2.5 family of models, making 2.5 Flash and 2.5 Pro stable and generally available while also unveiling a preview of Gemini 2.5 Flash-Lite — the most cost-efficient and fastest 2.5 model yet.

Gemini 2.5 Flash and Pro now stable for production use

Following extensive feedback from developers and businesses, Google is now making Gemini 2.5 Flash and 2.5 Pro broadly available for production applications. The two models have already been in use by companies like Snap, SmartBear, Spline, and Rooms in their applications over the last few weeks. The stable release means businesses can rely on these models for their workloads, from enterprise automation to developing sophisticated consumer apps.

Gemini 2.5 Flash-Lite preview promises speed and cost-efficiency

Google is also previewing Gemini 2.5 Flash-Lite — a lightweight, cost-effective and fast addition to its 2.5 family. The preview lets developers trial the model’s capabilities and provide feedback. According to Google, 2.5 Flash-Lite performs strongly across a range of coding, math, science, reasoning and multimodal tasks. It is designed for latency-sensitive workloads, delivering faster responses while consuming less compute — making it a viable option for large-scale applications where cost and speed are key consideration factors.

All-around capabilities with 1M context length

Despite its lightweight design, Gemini 2.5 Flash-Lite comes with all the capabilities that make the 2.5 family powerful. It supports extensive context — up to 1 million tokens — allowing it to handle large documents, conversations and codebases. It can connect to Google Search and code execution tools, handle multimodal inputs and respond accurately across a range of tasks.

Availability across platforms

Developers can now find stable versions of Gemini 2.5 Flash and Pro in Google AI Studio, Vertex AI and the Gemini app. The preview of Gemini 2.5 Flash-Lite is available in Google AI Studio and Vertex AI. Furthermore, custom versions of 2.5 Flash-Lite and 2.5 Flash have been integrated into Google Search, extending their capabilities to more applications.