Google for Developers: Building with Gemma 4

Gemma 4 is the most capable model family you can run on your hardware. They complement our Gemini models, giving our developer community the industry's most powerful combination of both open and proprietary tools. Watch for inspiration for your next Gemma build, or even just to get to know Reachy the offline robot (running LFG-3 on a Reachy Mini and NVIDIA DGX Spark).



Learn more about Gemma here: https://goo.gle/4dxj30m, and share what you’ll build with an AI model that runs directly on-device in the comments below!





🔗Resources:



Community Builds Featured in this Video

*Wearable Robotic Arm → https://goo.gle/4rVOib9

*PenguinAgent → https://goo.gle/3VVfTNO

*CodeBuddy (Offline Coding) → https://goo.gle/4y9tRsM



Models & Open Source Attributions

*Custom Model: LFG-3 → https://goo.gle/3UbydBQ

*Speech Recognition: LFG-3 fuses NVIDIA Parakeet → https://goo.gle/3TTzn4T, licensed under CC BY 4.0 → https://goo.gle/4ic5Tby

*Text-to-Speech: This project uses Kyutai TTS 1.6B → https://goo.gle/4hBVrtR, licensed under CC BY 4.0 → https://goo.gle/4ic5Tby





Subscribe to Google for Developers → https://goo.gle/developers







Speakers: Glenn Cameron

Products Mentioned: Gemma