Gemini Presents 2.5 Flash Model For Developers

Gemini presents the latest model of reasoning (doc. Google).

JAKARTA Google started rolling out the initial version of Gemini 2.5 Flash on April 17. The Gemini model is available in the form of a preview at Gemini API for app developers.

Developers can access it through Google AI Studio and Vertex AI, according to a Google report. Compared to the 2.0 Flash which is also popular, this device brings a bigger improvement in terms of reasoning.

Google also says that they prioritize speed and cost in Gemini's latest Flash version. That way, developers can optimize their work by activating thought mode.

"Our new 2.5 Flash MOdel has an outstanding performance-to-cost ratio, which places it on the Pareto border. This is our first fully hybrid reasoning model," Google said, quoted on Saturday, April 19, 2025.

In addition, this latest reasoning model is also claimed to be able to help developers balance the quality, costs, and latency of the applications they develop. Even if the thought mode is disabled, 2.5 Flashes will continue to work optimally.

"In fact, by disabling the thought, the developer can maintain a speed of 2.0 Flashes and improve performance. We can't wait to see how you use Gemini 2.5 Flashes in your app," Google said.

This latest reasoning model is not only released for developers, but also for all Gemini users. Later, this artificial intelligence (AI) model can be used with third-party applications such as Canvas to perfect documents or codes.