On December 5, 2025, Google published a roundup of AI updates announced in November across its products, research, and infrastructure. The company highlighted new Gemini 3 capabilities, image generation tools, navigation features, science and education initiatives, and large-scale AI and cloud investments in Texas and other regions.
Key Details
Google introduced Gemini 3 as its latest multimodal AI model for learning, planning, and problem solving. Gemini 3 is available through the Gemini app and a web interface, with additional technical and product details provided in a dedicated announcement hub.
Gemini 3 powers Google's new AI Mode in Search, which offers dynamic visual layouts, interactive tools, and simulations. Subscribers to Google AI Pro and Ultra in nearly 120 regions can select Gemini 3 Pro in AI Mode in Search.
Google also launched Nano Banana Pro, an image generation and editing model built on Gemini 3. Users can choose between the original Nano Banana for quick edits or Nano Banana Pro for higher detail and more complex image tasks.
The company introduced Google Antigravity, an agentic development platform focused on AI-powered coding and autonomous task execution. Antigravity is available as a public preview through a dedicated site for developers.
Google Maps will add Gemini-based, voice-driven navigation that supports conversational queries, traffic reports, and route suggestions while driving. Landmark-based navigation is rolling out in the United States on Android and iOS, using nearby businesses and locations for guidance, as detailed in Google's Gemini navigation and landmark features and global navigation rollout updates.
Gemini capabilities are also arriving in Android Auto, which runs in more than 250 million cars globally. Drivers can use natural language to add stops, send messages, manage media, and generate ideas while on the road through Android Auto with Gemini.
Background Context
Google stated that it has invested in machine learning and AI for more than 20 years to support its products. The November roundup highlighted work across consumer tools, developer platforms, scientific research, and infrastructure.
Google DeepMind introduced SIMA 2, a game-playing AI agent that operates in virtual 3D environments. SIMA 2 integrates Gemini features and can follow language instructions, reason about goals, converse, and improve performance over time. The company also released WeatherNext 2, a forecasting model it reports is eight times faster and supports hourly resolution.
Google marked five years since AlphaFold 2 solved the protein structure prediction problem and began driving new biological research. The roundup noted AlphaFold's 2024 Nobel Prize recognition and described how the system has supported new research directions worldwide.
New AI planning tools in Search let users describe trips and build itineraries using the Canvas feature in AI Mode, as outlined in Google's travel planning with Canvas update. Google also added conversational AI shopping features in Search and Gemini, including stock checks and automated checkout for eligible merchants, detailed in its holiday AI shopping features announcement.
At its AI for Learning Forum in London, Google announced $30 million in new funding focused on learning initiatives. The company expanded NotebookLM with tools to generate flashcards and quizzes to support day-to-day studying, as described in its NotebookLM updates. Gemini Live received improvements for more natural, two-way audio conversations and adjustable speech speed for accessibility and pacing, covered in the Gemini Live audio updates. These moves sit within Google's broader AI learning commitments and education initiatives.
Google announced a $40 billion investment program in Texas for AI and cloud infrastructure at an event in Midlothian, described in its Texas AI and cloud investment update. CEO Sundar Pichai also detailed related investments across the United States, Europe, Africa, and the Asia Pacific region, building on initiatives like Investing in America 2025, Europe's AI plans, and AI-ready data efforts in Africa. Google also reported a workforce initiative to train 100,000 electrical workers and create 30,000 new apprentices, as outlined in its electrical workers training program.
Source Citations
- Google November 2025 AI roundup overview: The latest AI news we announced in November.
- Gemini 3 model and app access: Gemini 3, Gemini 3 in the Gemini app.
- Gemini 3 for developers and AI Mode in Search: Gemini 3 for developers, Gemini 3 in Search AI Mode.
- Nano Banana Pro image model and guidance: Nano Banana Pro, Developer image documentation, Prompting tips.
- Google Antigravity platform: Developer announcement, Public preview.
- Maps and Android Auto with Gemini: Gemini navigation and landmark features, Global navigation rollout, Android Auto with Gemini.
- Research and science updates: SIMA 2, WeatherNext 2, AlphaFold impact.
- AI travel planning and shopping: Travel planning with Canvas, Holiday AI shopping features.
- Education and NotebookLM updates: AI learning commitments, AI and learning, NotebookLM quizzes and flashcards.
- Gemini Live audio improvements: Gemini Live updates.
- Infrastructure, regional investments, and workforce training: Texas AI and cloud investment, Investing in America 2025, Europe AI plans, AI-ready data in Africa, DeepMind in Singapore, Electrical workers training.
.svg)
.svg)
.svg)