Google capped the year with a broad slate of AI upgrades aimed at consumers, developers and enterprises. The company introduced Gemini 3 Flash as the default model in the Gemini app and AI Mode in Search, promising faster reasoning at lower cost while extending availability through Vertex AI and its new agentic platform, Antigravity. To bolster trust, Google added video verification in Gemini using SynthID watermarks across audio and visuals. Labs project “Disco” and its GenTabs feature target complex browsing by synthesizing tabs and chats into interactive tools. Voice capabilities advanced with Gemini 2.5 Flash Native Audio and a beta for live speech translation in Google Translate. A new Gemini Deep Research agent arrived via the Interactions API, paired with the open-source DeepSearchQA benchmark. On the commerce front, U.S. shoppers can now generate full-body try-ons from a selfie. Google also expanded access to Pro-tier models in Search, and highlighted engagement features with YouTube’s first personalized Recap, new Google Photos Recap controls and its Year in Search review, underscoring a shift toward conversational queries.
Related articles:
– Vertex AI: Build, deploy, and scale AI with Google Cloud
– Google AI Research: Publications and highlights
– Use conversation mode in Google Translate
– What’s new in Google Photos
