Google I/O 2025: Gemini 2.5 Flash and Gemma 3n Empower Developers with Enhanced AI Tools

On May 20, 2025, Google kicked off I/O with a wave of updates focused on developers building with AI. The company launched Gemini 2.5 Flash, a leaner version of its flagship model, built for speed. It improves code generation and reasoning without bloating compute costs.

Joining it is Gemma 3n, a lightweight multimodal model that runs on phones and laptops. It can process text, audio, images, and video. Developers can access it through Google AI Studio or push it to edge devices with Google AI Edge.

Also read: Google I/O 2025: Top 10 Highlights of Google’s New Announcements

To give developers more control, Google added “thought summaries” across its 2.5 models. These show how the model forms responses. Google also previewed “thinking budgets” for Gemini 2.5 Pro, letting developers cap resource usage per task.

The updates show Google isn’t just scaling up its AI—it’s trying to hand developers the tools to make it usable, fast, and efficient across hardware.

While Gemini Advanced and other consumer-facing features made headlines, Google I/O 2025 was a strong signal that it wants developers building AI-native apps—not just using them.