Google has just introduced a new member to its Gemini 2.5 AI family—theGemini 2.5 Flash. It is claimed to be built to be speedy, efficient, and light on resources, making it ideal for real-time conversations, instant summarisation, and large-scale interactions where quick thinking is key.

In a blog post announcing the launch, the tech giant said that the Flash model is now available on both Google AI Studio and Vertex AI. That means developers and businesses can start building applications and AI agents with it straight away.

Unlike its more advanced sibling,Gemini 2.5 Pro — which is also now live on Vertex AI — Flash is all about performance without the heavy lifting. Where the Pro version handles deep analysis and complex decision-making, Flash focuses on speed and affordability, making it perfect for use cases like chatbots, virtual assistants, and other tools that need to respond quickly and efficiently, said Google.

Google is calling Flash a “workhorse model”, and it comes with built-in “dynamic and controllable reasoning”. What that means is developers can tweak how much processing time a query gets, depending on how tricky it is — giving them more control over how fast or detailed the AI’s response needs to be.

For those unsure about which model to use, Google is also testing a newModel Optimiser tool within Vertex AI. It helps pick the best model for a task based on factors like quality and cost, so users do not have to guess their way through.

And there is more: the company is adding a Live API  powered by Gemini 2.5 Pro to help build even smarter, more responsive AI agents. This new tool can process streaming audio, video, and text in real-time. It also supports longer sessions (over 30 minutes), multilingual audio output, time-stamped transcripts, and integrations with other tools.