Google has energized the tech world with two major announcements aimed at solidifying its leadership in the AI ecosystem. While the company officially unveiled the Gemma 4 family—the newest generation of its open-source AI models—it also released a comprehensive update for its video creation platform, Google Vids, taking productivity to a whole new level.
Gemma 4: A New Era in the Open-Source World
Developed by Google DeepMind, Gemma 4 is designed to challenge commercial models despite its “open-source” tag. Sharing the same technological foundation as Gemini 3, the Gemma 4 family arrives in four distinct model sizes: Effective 2B, Effective 4B, 26B (Mixture of Experts), and 31B (Dense).
Key highlights of the new model family include:
- Massive Context Window: The smaller models offer a 128K context window, while the larger versions scale up to 256K, allowing them to process thousands of pages of documents or massive code blocks in a single go.
- Multimodal Capabilities: Gemma 4 natively understands images and audio data alongside text. It shows industry-leading performance in handwriting recognition, chart analysis, and voice-based Q&A.
- Mobile and IoT Optimization: The Effective 2B and 4B models are optimized to run offline on smartphones and devices like Raspberry Pi or NVIDIA Jetson without needing an internet connection.
- Global Language Support: Trained in over 140 languages, the model provides significantly more natural results in multilingual workflows.
Google Vids: One-Click Video with Veo 3.1
Google Vids, the video-centric app for Google Workspace, has been transformed into a professional editing studio with its latest update. At the center of this update is Veo 3.1, Google’s most advanced video generation model.
New features included in the update:
- Free Video Generation: Free Google accounts can now generate up to 10 AI videos per month (8 seconds each) using text prompts or image uploads.
- Advanced AI Avatars: Users can customize the appearance, clothing, and backgrounds of their avatars. The new “Direct Avatars” feature allows AI characters to interact with objects, a game-changer for product demos and tutorials.
- Lyria 3 Integration: Paid plan users can now generate high-quality AI music for their videos, powered by Lyria 3.
- Chrome Screen Recording: A new extension allows users to capture screen recordings instantly and integrate them into Vids projects without opening the app.
New Tools for Developers
Google also announced the Agent Development Kit (ADK), a framework designed to work seamlessly with Gemma 4. This toolkit will allow developers to rapidly build autonomous AI agents capable of making decisions, calling functions, and managing complex logical flows.
{{user}} {{datetime}}
{{text}}