LiteRT is the universal framework for on-device AI. The production stack delivers 1.4x faster cross-platform GPU performance, streamlined NPU acceleration, and superior GenAI support for open models like Gemma.

GoogleDevs' platform is a central hub for developers interested in Google technologies, APIs, and developer tools. Through articles, tutorials, and documentation, GoogleDevs offers insights into building applications using Google Cloud Platform, Android, Chrome, and other Google services. Developers can learn about cloud computing, machine learning, and mobile app development with Google's developer tools and platforms.

Google Developers

LiteRT, Google's on-device AI inference framework successor to TensorFlow Lite, has graduated its advanced acceleration capabilities into full production. Key highlights include cross-platform GPU support (Android, iOS, macOS, Windows, Linux, Web) via the ML Drift engine delivering 1.4x faster performance over legacy TFLite GPU, streamlined NPU acceleration with production-ready integrations for MediaTek and Qualcomm (up to 100x faster than CPU), and GenAI support for open models like Gemma 3. The framework supports model conversion from PyTorch, TensorFlow, and JAX, and introduces asynchronous execution and zero-copy buffer interoperability for up to 2x latency reduction in real-time use cases.

LiteRT: The Universal Framework for On-Device AI

High-performance cross-platform GPU acceleration

Streamlined NPU integration with peak performance

Reliability and compatibility you can trust