---
markdown_generated_at: "2026-11-10"
title: "Qualcomm® AI Runtime SDK (QAIRT)"
url: "https://www.qualcomm.com/developer/software/qualcomm-ai-runtime-sdk-qairt"
---

# Qualcomm® AI Runtime SDK (QAIRT)

![](https://s7d1.scene7.com/is/image/dmqualcommprod/foundries-factory-blank-desktop)

[Get software](https://softwarecenter.qualcomm.com/api/download/software/sdks/Qualcomm_AI_Runtime_Community/All/2.44.0.260225/v2.44.0.260225.zip)

[See all versions](https://softwarecenter.qualcomm.com/catalog/item/Qualcomm_AI_Runtime_Community)

[Get Started](https://docs.qualcomm.com/nav/home?product=924033590759186372)

![chip image](https://s7d1.scene7.com/is/image/dmqualcommprod/QAIRT-icon-x2?fmt=webp-alpha&qlt=85)

## Build Once, Deploy Everywhere with Qualcomm AI Runtime

Qualcomm® AI Runtime SDK (QAIRT) is a suite of tools and runtimes to build, optimize, and run ML models generated by TensorFlow, PyTorch, and ONNX across the CPU, GPU, NPU, and [Qualcomm Sensing Hub](https://docs.qualcomm.com/doc/80-PV086-5P/topic/qualcomm-sensing-hub.html?product=1601111739937065) on mobile phones, tablets, Windows PCs, automotive systems, IoT devices, and embedded platforms powered by Snapdragon and Dragonwing platforms.

Product license agreement

### Featured Resources

[QAIRT Overview](https://docs.qualcomm.com/doc/80-63442-10/topic/general_overview.html)

[Discord](https://discord.com/invite/qualcommdevelopernetwork)

## Use QAIRT Directly or Through Popular ML Framework Runtimes

There are two ways to use QAIRT:

![](https://s7d1.scene7.com/is/image/dmqualcommprod/qairt-2x-7-20)

## Use QAIRT Directly

Build your own end-to-end native pipelines for building, deploying, debugging, and benchmarking ML models where you need to minimize inference latency in real-time applications, or optimize for power efficiency.

The QAIRT runtime API for C/C++ and Python provides various levels of control over model execution. Start by simply loading and running models and then add granular performance and efficiency enhancements as you delve deeper into model graph optimization, hardware-specific operators, and quantization strategies (e.g., FP32/16, INT8).

QAIRT also includes [Qualcomm® Generative AI Inference Extensions (GENIE)](https://www.qualcomm.com/developer/software/gen-ai-inference-extensions), a software library specifically for deploying Gen AI models comprised of multiple binaries that require a specific order of execution.

For developers familiar with the Qualcomm® AI Stack, QAIRT combines features previously available on the Snapdragon Neural Processing Engine (SNPE) SDK and the Qualcomm AI Engine Direct (QNN) SDK into a single, comprehensive solution.

## Use Popular ML Framework

Use popular ML framework runtimes built on QAIRT including:

- LiteRT (TensorFlow Delegate)
- [ExecuTorch (PyTorch Delegate)](https://docs.pytorch.org/executorch/stable/backends-qualcomm.html)
- [ONNX RT (ONNX Execution Provider)](https://github.com/onnxruntime/onnxruntime-qnn)

Each runtime abstracts the model execution details on Snapdragon® or Qualcomm Dragonwing™ processors. This approach lets you work with your models in your preferred framework without having to work directly with QAIRT but limits you to the features supported by your chosen framework. [See QAIRT Overview for more information.](https://docs.qualcomm.com/doc/80-63442-10/topic/general_overview.html)

For Windows developers adding AI to a Windows on Snapdragon applications, find more detailed information [here.](https://www.qualcomm.com/developer/windows-on-snapdragon/windows-on-snapdragon-ai)

For IoT and embedded developers, find more detailed information [here.](https://www.qualcomm.com/developer/iot)

More information on the Qualcomm AI Stack and ecosystem can be found [here.](https://www.qualcomm.com/developer/artificial-intelligence)

## Benefits

### Flexible Levels of Control

Work at a level that is right for you. The QAIRT runtime API provides various levels of control that you can choose for performance and optimization based on your requirements.

### Flexible Workflows

Download a single SDK that supports ML model pipelines and runtime integrations across multiple ML frameworks or use an ML framework that is built on QAIRT but abstracts away its details.

### Stability Across Releases

Simplify your testing and reduce churn with built-in Op versioning. Compile and run AI models through stable interfaces across SDK updates and new hardware releases, so you decide when to adopt newer versions.

### Compile Once, Test Once

API versioning ensures application binary interface (ABI) compatibility with different library versions. Query for a specific interface at runtime, so you don’t have to recompile when you update the QAIRT library or you can run on other supported devices powered by Snapdragon or Dragonwing processors.

## Blog

- OnQ Blog
- [AI on Snapdragon is AI that moves with you](https://www.qualcomm.com/news/onq/2026/11/ai-on-snapdragon-agentic-age)
- Nov 9, 2026 | 6 min read
- Tags: [Snapdragon](https://www.qualcomm.com/news/onq?tags=Snapdragon), [AI](https://www.qualcomm.com/news/onq?tags=AI), [Agentic AI](https://www.qualcomm.com/news/onq?tags=Agentic%20AI)

- OnQ Blog
- [How Qualcomm Dragonwing is putting industrial and physical AI to work](https://www.qualcomm.com/news/onq/2026/11/qualcomm-dragonwing-industrial-physical-ai)
- Nov 6, 2026 | 5 min read
- Tags: [Dragonwing](https://www.qualcomm.com/news/onq?tags=Dragonwing), [Industrial IoT](https://www.qualcomm.com/news/onq?tags=Industrial%20IoT), [Physical AI](https://www.qualcomm.com/news/onq?tags=Physical%20AI)

- OnQ Blog
- [Qualcomm HBC demonstrates AI inference leadership with validated silicon](https://www.qualcomm.com/news/onq/2026/11/qualcomm-hbc-validated-silicon-ai-inference)
- Nov 2, 2026 | 3 min read
- Tags: [Dragonfly](https://www.qualcomm.com/news/onq?tags=Dragonfly), [HBC](https://www.qualcomm.com/news/onq?tags=HBC), [AI Inference](https://www.qualcomm.com/news/onq?tags=AI%20Inference)

- Developer Blog
- [How to write fast AArch64 SVE/SME Kernels with Ripple and LLVM](https://www.qualcomm.com/developer/blog/2026/10/ripple-llvm-aarch64-sve-sme-kernel-implementation)
- Oct 29, 2026
- Tags: [LLVM](https://www.qualcomm.com/developer/blog?tags=LLVM), [Open Source](https://www.qualcomm.com/developer/blog?tags=Open%20Source)

- Developer Blog
- [Same-day upstream Linux support now posted for Snapdragon 8 Elite Gen 6 and Snapdragon Elite Extreme Gen 6 mobile platforms](https://www.qualcomm.com/developer/blog/2026/10/same-day-snapdragon-8-elite-gen-6-upstream-linux-support)
- Oct 27, 2026
- Tags: [Adreno](https://www.qualcomm.com/developer/blog?tags=Adreno), [Linux](https://www.qualcomm.com/developer/blog?tags=Linux), [Snapdragon](https://www.qualcomm.com/developer/blog?tags=Snapdragon)

## Connect with our communities

[LinkedIn](https://www.linkedin.com/showcase/qualcomm-developer)

[See what's new](https://x.com/Qualcomm_Dev)

[YouTube](https://www.youtube.com/qualcommdev)

[GitHub](https://github.com/qualcomm)

![](https://s7d1.scene7.com/is/image/dmqualcommprod/discord-icon-1?$QC_Responsive$&fmt=png-alpha)

[Discord](https://discord.com/invite/qualcommdevelopernetwork)

[Visit the forum](https://mysupport.qualcomm.com/supportforums/s/)

![](https://s7d1.scene7.com/is/image/dmqualcommprod/ai-hub-bg-desktop?fmt=webp-alpha&qlt=85)