google-ai-edge/LiteRT-LM v0.10.1: Production Edge LLM Inference with Function Calling and NPU Acceleration

GitHub·medium signal

Google's LiteRT-LM shipped v0.10.1 on April 3 with production-ready inference for Gemma, Llama, Phi-4, and Qwen models across Android, iOS, Web, Desktop, and IoT devices. Key capabilities include hardware acceleration via GPU and NPU, multimodal vision/audio inputs, and function calling for agentic workflows on-device. This is the same framework powering on-device GenAI in Chrome, Chromebook Plus, and Pixel Watch — now available for any developer to deploy edge agents with tool-calling capabilities.