Google rolls out Gemini 3 Flash and expands AI features across apps in December update
Faster Gemini models, video verification, live translation and new developer tools underline push to move AI from research labs into everyday products
Google has outlined a broad set of artificial intelligence (AI) updates for December, led by the launch of Gemini 3 Flash and new capabilities across Search, Translate, Photos and YouTube.
The company said it has invested for more than 20 years in machine learning and AI research and is focused on taking frontier intelligence out of the lab and into users’ hands.
Gemini 3 Flash combines high speed with improved reasoning and is rolling out globally as the default model in the Gemini app and AI Mode in Search. Google said the rollout also supports developers via the application programming interface, the Antigravity agentic platform and enterprise customers using Vertex AI.
The Gemini app now supports video uploads of up to 100 megabytes or 90 seconds for verification. Google said it is using SynthID watermarks to identify AI-generated audio and visual segments within uploaded content.
Additional updates include the launch of Disco and the GenTabs experiment, designed to synthesise open browser tabs and chat history. Google also upgraded Gemini 2.5 Flash Native Audio across AI Studio, Vertex AI, Gemini Live and Search Live.
In translation, Google introduced a live speech translation beta in the Google Translate app, supporting more than 70 languages. The feature is aimed at enabling real-time spoken translation during conversations.
For developers, Google unveiled a new Gemini Deep Research agent through the Interactions application programming interface and open-sourced the DeepSearchQA benchmark to support research and evaluation.
Consumer-facing features also expanded. US shoppers can now use Nano Banana virtual try-on, which allows clothing visualisation from a single selfie. Gemini 3 Pro and Nano Banana Pro are available in AI Mode in English across nearly 120 countries and territories, with higher usage limits and expanded access for Google AI Pro and Ultra subscribers in the US.
Separately, YouTube released its annual top trends alongside a first-ever personalised Recap feature, while Google Photos updated Recap controls and sharing options.
Google said the updates reflect its strategy of embedding AI across products rather than treating it as a standalone tool.
The Recap
- Google summarised its December AI product and model releases.
- Gemini video uploads accepted up to 100 MB or 90 seconds.
- Gemini 3 Flash rolls out as default model in apps.
