Google I/O 2025 Highlights: Gemini Models, AI Mode in Search, and Multimodal AI

At Google I/O 2025, Google highlighted major advancements in its Gemini AI models, including the release of Gemini 2.5, which powers a new AI Mode in Google Search. This AI Mode enables users to perform complex, multimodal searches that break down intricate queries into subtopics, delivering detailed and comprehensive answers by running multiple simultaneous searches on their behalf. Initially launched in the U.S., AI Mode has expanded to Singapore, allowing users to ask longer, more nuanced questions and receive unified, in-depth responses that combine advanced AI reasoning with extensive web content.

Key highlights include:

• Gemini 2.5 Models: The latest iteration of Google's language models, offering improved intelligence and multimodal capabilities, underpinning AI Mode in Search and other generative AI tools.

• AI Mode in Search: A new search experience that goes beyond traditional search by using a "query fan-out" technique to dissect complex questions into smaller parts, enabling deeper exploration and richer answers. It supports follow-up questions and integrates helpful web links for further exploration.

• Multimodal AI: AI Mode supports multimodal queries, meaning users can combine text with other input types (such as images) to get more contextual and detailed results, enhancing the search experience.

• Global Rollout: After initial testing in the U.S., AI Mode is now available in Singapore, reflecting Google's strategy to incrementally introduce AI-powered tools internationally.

• Broader AI Ecosystem: Alongside Search, Google announced upgrades to its Gemini app and generative AI tools, emphasizing AI’s growing role in simplifying tasks, boosting creativity, and enhancing productivity across products.

These innovations mark a significant step toward more intelligent, personalized, and action-oriented AI experiences integrated into everyday technology.