News

Google Expands Gemini 2.5 AI Model Line with Flash-Lite Preview

• By John K. Waters
• 06/17/2025

Google has announced general availability of its Gemini 2.5 Flash and Pro models and introduced a new addition to the lineup, Gemini 2.5 Flash-Lite, now in preview. The company said the new model is designed to be its fastest and most cost-efficient in the 2.5 family to date.

The release was announced in a blog post by Tulsee Doshi, Senior Director of Product Management, who noted that Gemini 2.5 models are designed as “hybrid reasoning models” optimized for both performance and speed.

“Thanks to all of your feedback, today we’re releasing stable versions of 2.5 Flash and Pro, so you can build production applications with confidence,” Doshi wrote. Companies such as Spline, Rooms, Snap, and SmartBear have already been using the models in production.

Flash-Lite Optimized for Latency-Sensitive Workloads

The new Gemini 2.5 Flash-Lite model is targeted at high-volume, low-latency tasks such as translation and classification. According to Google, it performs better than Gemini 2.0 Flash-Lite across coding, math, science, reasoning, and multimodal benchmarks, with lower latency compared to both 2.0 Flash and Flash-Lite on a wide range of prompts.

Flash-Lite includes the standard 2.5 capabilities: flexible reasoning budgets, tool use (such as Google Search and code execution), multimodal input support, and a 1 million-token context window.

Read more technical details in the Gemini 2.5 Technical Report.

Deployment Availability

The Flash-Lite preview is currently available through Google AI Studio and Vertex AI. The stable versions of Gemini 2.5 Flash and Pro are also accessible via those platforms, as well as through the Gemini app. Custom variants of Flash and Flash-Lite have been integrated into Google Search.

“We can’t wait to see what you continue to build with Gemini 2.5,” Doshi said.