Google Search Integrates Gemini 2.5 Pro, Expanding AI Capabilities with 1 Million Token Context Window

Google Search is significantly upgrading its AI capabilities with the integration of the Gemini 2.5 Pro model, offering users advanced reasoning and multimodal query options. These enhancements, detailed in a recent post by Google DeepMind's Senior AI Relation Engineer Philipp Schmid, are rolling out to Google AI Pro and AI Ultra subscribers, initially in regions like the US and India. The update aims to transform how users interact with search, enabling more complex and conversational queries.

The newly integrated system supports a wide array of query methods, including text, voice, images, and PDF uploads, allowing for a more versatile user experience. A key feature, "query fan-out," enables the AI to simultaneously search multiple subtopics, providing comprehensive results. Users can also resume previous searches through an AI Mode history, fostering a continuous and refined research process.

At the core of these advancements is the Gemini 2.5 Pro model, known for its advanced reasoning and native multimodal capabilities. This model can process and understand diverse data types, including text, audio, images, and video, significantly broadening the scope of searchable content. Furthermore, it boasts a substantial 1-million token context window, allowing the AI to maintain extensive context over prolonged and intricate interactions.

Leveraging Gemini 2.5 Pro, Google is also introducing "Deep Search" within AI Mode, designed to conduct hundreds of background searches and synthesize information into comprehensive, cited reports. Additionally, an "agentic" AI-powered calling feature is rolling out to all US users, with higher limits for subscribers, enabling the AI to contact local businesses for information like pricing and availability. This marks a shift towards more task-oriented and autonomous search tools.

Philipp Schmid, a prominent figure in AI development and formerly a Technical Lead at Hugging Face, has been instrumental in making advanced AI accessible and has discussed building agentic AI systems with Google Gemini. The enhanced AI Mode, featuring Gemini 2.5 Pro and Deep Search, is available for Google AI Pro and AI Ultra subscribers who have opted into the AI Mode experiment in Google Labs, with wider accessibility planned. This strategic move positions Google Search as an increasingly intelligent and interactive reasoning assistant.