Google launches stable Gemini 2.5 Flash and Pro models and previews the new 2.5 Flash-Lite, offering faster, cost-efficient AI performance. Now available via AI Studio, Vertex AI, Gemini app, and integrated into Google Search.
Google has officially launched the stable versions of its Gemini 2.5 Flash and Pro models, further expanding its next-gen AI lineup. In addition, the company has introduced a preview of Gemini 2.5 Flash-Lite — its most efficient and fastest model in the 2.5 family so far.
Designed to balance high performance with cost and speed, the Gemini 2.5 models are built around hybrid reasoning capabilities. With the stable rollout of Gemini 2.5 Pro and Flash, developers can now confidently use them to build and scale production-ready applications. These models have already seen early real-world deployment by developers and companies like Spline, Rooms, Snap, and SmartBear, who helped refine the offerings through their feedback.
The spotlight now also turns to the newly previewed Gemini 2.5 Flash-Lite. Built for high-volume, latency-sensitive tasks like classification and translation, Flash-Lite is both faster and more cost-effective than its predecessors — 2.0 Flash and Flash-Lite. It also delivers better performance across areas such as coding, math, science, reasoning, and multimodal benchmarks. The model supports key Gemini features, including multimodal input, tool integration (like Google Search and code execution), and a 1 million-token context length.
Also Read: Google Gemini Rolls Out Special Feature to Automate Routine Tasks for Users
Google is making these models widely accessible. The stable Gemini 2.5 Pro and Flash, along with the Flash-Lite preview, are available through Google AI Studio, Vertex AI, and the Gemini app. Additionally, customized versions of Flash and Flash-Lite are already embedded in Google Search, enhancing user experiences across the platform.
