[Google Launches FunctionGemma 270M Edge AI Model] According to Google's official blog, Google has released FunctionGemma, specifically fine-tuning Gemma 3 270M as a function-calling model for local/offline agent scenarios. Its features include unified chat and tool execution, support for custom fine-tuning (improving Mobile Actions accuracy from 58% to 85%), miniaturization for edge devices (such as NVIDIA Jetson Nano and mobile phones), and optimization for JSON/multilingual input. Official resources include downloads on Hugging Face and Kaggle, as well as fine-tuning guides for Transformers, Unsloth, Keras, NeMo, and deployment instructions for LiteRT‑LM, vLLM, MLX, Llama.cpp, Ollama, Vertex AI, and LM Studio. Additionally, the Edge Gallery showcases multiple demos, datasets, and Colab resources.