Qualcomm

Qualcomm® AI Hub is making it easier for everyone to run AI models on-device!

Qualcomm® AI Hub Models showcases dozens of pre-optimized, ready-to-deploy AI models across mobile, PC, IoT and automotive devices. Browse the collection, review performance and select a model to download today.

To run your model on device check out the path that fits your workflow:

• GenieX (Developer Preview) — our on-device SDK for running any Gen AI model on Snapdragon across NPU, GPU, and CPU with a few lines of code. Leverage our QAIRT plugin for optimal NPU performance on Qualcomm AI Hub Models, and llama.cpp plugin for broad coverage of community GGUF models from Hugging Face.
• Qualcomm AI Hub Workbench — compile, run inference, profile, and quantize on cloud-hosted devices to iterate on model performance before you ship. Sign up today

Get started in 5 minutes on Windows, Android, or Linux via CLI, Python, Maven, Docker, or OpenAI-compatible APIs. Ship on-device with the runtime of your choice — GenieX, TensorFlow Lite, ONNX Runtime, or the Qualcomm® AI Engine Direct SDK.

Join our AI Hub Slack community to collaborate, ask questions and learn more about on-device AI. For questions or feedback please reach out to us.