# Compile and optimize model Source: [https://docs.qualcomm.com/doc/80-70014-15B/topic/compile-optimize.html](https://docs.qualcomm.com/doc/80-70014-15B/topic/compile-optimize.html) Users can take one of two paths available to compile and optimize their models. **AI Hub** [**AI Hub**](https://docs.qualcomm.com/doc/80-70014-15B/topic/ai-hub.html) Bring your own model or try preoptimized models on Snapdragon in < 5 min. **AI Software Stack** [**TFLite**](https://docs.qualcomm.com/doc/80-70014-15B/topic/tflite.html) Port your TFLite AI models to Snapdragon [**SNPE**](https://docs.qualcomm.com/doc/80-70014-15B/topic/snpe.html) Port your models using the all-in-one, easily customizable SNPE SDK [**QNN**](https://docs.qualcomm.com/doc/80-70014-15B/topic/qnn.html) Port your models using modular, low-level QNN APIs Last Published: Jan 21, 2026 [Previous Topic Overview](https://docs.qualcomm.com/bundle/publicresource/80-70014-15B/topics/offering.md) [Next Topic AI Hub](https://docs.qualcomm.com/bundle/publicresource/80-70014-15B/topics/ai-hub.md)