Google is releasing a new AI model designed to deliver strong performance with an emphasis on efficiency.
The model Gemini 2.5 Flash will soon be released on Vertex AI, Google’s AI development platform. The company says it offers “dynamic and controllable” computing, allowing developers to adjust processing times based on the complexity of the query.
“(You can tune it) to check the speed, accuracy and cost balance for your specific needs,” Google wrote in a blog post provided to TechCrunch. “This flexibility is key to optimizing flash performance in large amounts of cost-sensitive applications.”
The Gemini 2.5 flash arrives and costs for flagship AI models continue to rise. Low-cost performance models like the 2.5 Flash offer attractive alternatives to costly top-notch options at the expense of some degree of accuracy.
Gemini 2.5 Flash is a “inference” model along the lines of Openai’s O3-Mini and Deepseek’s R1. This means that it takes a little time to answer the question to fact-check itself.
Google says 2.5 flash is perfect for “volumes” and “real-time” applications such as customer service and document analytics.
“This flagship model is especially optimized due to low latency and reduced costs,” Google said in a blog post. “This is the ideal engine for responsive virtual assistants and real-time summarization tools where efficiency at scale is critical.”
Google didn’t publish safety or technical reports for Gemini 2.5 Flash, making it difficult to see where the models are good and lacking. The company previously told TechCrunch it had not released a report on a model that it would be considered “experimental.”
Google announced Wednesday it plans to bring Gemini models like the 2.5 Flash into on-premises environments starting in the third quarter. The company’s Gemini model is available in Google Distributed Cloud (GDC), Google’s ONPREM solution for clients with strict data governance requirements. Google says it is working with Nvidia to bring the Gemini model to the GDC-compliant Nvidia Blackwell system.