Google on Tuesday revealed new Android development tools, a new mobile AI architecture, and an expanded developer community. The announcements accompanied the unveiling of an AI Mode for Google Search at the Google I/O keynote in Mountain View, California.

Google’s Gemini Diffusion demo didn’t get much airtime at I/O, but its blazing speed—and potential for coding—has AI insiders speculating about a shift in the model wars.

Google says the release version of 2.5 Flash is better at reasoning, coding, and multimodality, but it uses 20–30 percent fewer tokens than the preview version. This edition is now live in Vertex AI, AI Studio, and the Gemini app. It will be made the default model in early June.

Gemini Diffusion is a new AI model by Google DeepMind that transforms jumbled inputs into coherent text or code at record speed. It is currently in the testing phase.

On Tuesday at Google I/O 2025, the company announced Deep Think, an “enhanced” reasoning mode for its flagship Gemini 2.5 Pro model. Deep Think allows the model to consider multiple answers to questions before responding, boosting its performance on certain benchmarks.

Google’s AI models are learning to reason, wield agency, and build virtual models of the real world. The company’s AI lead, Demis Hassabis, says all this—and more—will be needed for true AGI.

Gemini 2.5 Flash plays the role of the default model in the Gemini chatbot now. It's supposed to be the fast, cost-efficient model for daily use. Google says it’s better than its predecessors, like Gemini 2.0 Flash, in terms of understanding images and text while still being much cheaper to run.

Google’s AI models have a secret ingredient that’s giving the company a leg up on competitors like OpenAI and Anthropic. That ingredient is your data, and it’s only just scratched the surface in terms of how it can use your information to “personalize” Gemini’s responses.