Google has announced the release of a new artificial intelligence model, Gemini 2.5 Flash, which offers enhanced capabilities for developers.
This model is the first hybrid system that allows you to customize the level of "thinking" for the optimal balance between cost, speed and quality. It is already available in beta via Gemini API, AI Studio and Vertex AI.
Gemini 2.5 Flash builds on the previous version of 2.0 Flash, maintaining its speed and low cost, but significantly improving the ability to perform logical reasoning. Developers can adjust the computational intensity and set a budget for each query, allowing them to effectively control the performance of the model.
Google also shared test results that demonstrate the versatility of the Gemini 2.5 Flash across math, science, logic, and code generation tasks. While it doesn't outperform its competitors in every way, it offers the best balance of speed, cost, and quality.
Additionally, the model is available in the Gemini app, where users can try out new features such as Canvas, a space for collaborative work on text and code.