> For the complete documentation index, see [llms.txt](https://docs.roboflow.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.roboflow.com/models/model-weights/download-roboflow-model-weights.md).

# Download Model Weights

To run your Roboflow models on your own hardware, you can either use Roboflow Inference (the recommended, automatic method) or manually download Model Weights (for specific edge cases).

### Automatic Download via Roboflow Inference (Recommended)

[Roboflow Inference](https://docs.roboflow.com/deployment/self-hosted/self-hosted) is our open-source, scalable system for running models locally on CPU and GPU devices.

**This is the fastest and most reliable way to get started**. When you use Inference, you don’t need to manage files or versioning; Roboflow Inference automatically fetches and caches your model weights the first time you run your code.

* **How it works**: On your first inference request, the weights are downloaded from Roboflow’s servers and stored locally. All future predictions use this local cache - images are not sent to the cloud.
* **Deployment options**:
  * [Workflows](https://docs.roboflow.com/workflows/deploy/deploy-a-workflow): For production-ready multi-step computer vision workflows
  * [Python inference SDK](https://docs.roboflow.com/deployment/self-hosted/self-hosted#run-a-model): For Python integration

### Manual Model Weights Download

Sometimes you may need the raw weights file (ex: a PyTorch `.pt` file) to run on devices Roboflow does not yet natively support, such as custom Android implementations. The formats you can download depend on the model architecture.

See the [Supported Models table](/models/supported-models.md) for weights download compatibility.

{% hint style="warning" %}
**Premium Feature**: Manual weights download is only available for paid users on Core plans and certain Enterprise customers. Read more on our [Pricing page](https://roboflow.com/pricing).
{% endhint %}

#### Export Formats

ONNX is the recommended export format. It is available for the widest range of architectures, and you can convert an ONNX file to run on TensorRT, OpenVINO, CoreML, or any other ONNX-compatible runtime. Roboflow Inference also runs models with ONNX Runtime.

<table data-search="false"><thead><tr><th>Format</th><th>Typical use</th><th>Available for</th></tr></thead><tbody><tr><td>ONNX</td><td>Cross-platform runtimes (ex: TensorRT, OpenVINO)</td><td>Roboflow 3.0, YOLOv8, YOLOv12, YOLO26, YOLO-NAS, RF-DETR, and classification models</td></tr><tr><td>PyTorch (.pt)</td><td>Custom PyTorch pipelines</td><td>Roboflow 3.0, YOLOv8, YOLOv12, YOLO26, YOLO-NAS, RF-DETR, and classification models</td></tr><tr><td>CoreML</td><td><a href="https://docs.roboflow.com/deployment/self-hosted/sdks/ios-sdk">iOS and Apple devices</a></td><td>Roboflow 3.0, YOLOv8, YOLOv12, YOLO26, RF-DETR, and classification models</td></tr><tr><td>TFLite (int8)</td><td><a href="https://docs.roboflow.com/deployment/self-hosted/sdks/openmv">Microcontrollers (ex: OpenMV)</a></td><td>Roboflow 3.0, YOLOv8, YOLOv12, and YOLO26</td></tr><tr><td>TensorFlow.js</td><td><a href="https://docs.roboflow.com/deployment/self-hosted/sdks/web-browser">Web browsers</a></td><td>Roboflow 3.0 and RF-DETR object detection models</td></tr><tr><td>Luxonis OAK / DepthAI</td><td><a href="https://docs.roboflow.com/deployment/self-hosted/sdks/luxonis-oak">OAK cameras</a></td><td>Roboflow 3.0 and YOLOv8 object detection models</td></tr></tbody></table>

* Instance segmentation, classification, and keypoint detection models export to ONNX. CoreML is also available for instance segmentation and classification models.
* RF-DETR models can also export to TFLite, TensorRT, and ExecuTorch with the open source [rf-detr library](https://github.com/roboflow/rf-detr) (experimental).
* RF-DETR NAS and SAM3 models run only through Roboflow Inference and do not support weight download.
* Fine-tuned multimodal models (ex: PaliGemma, Florence-2, Qwen) download as their Hugging Face output directory (safetensors) without conversion.

#### Method A: Roboflow Platform

Navigate to the model within your Project. If your plan allows, clicking the "Download Weights" button will allow you to download the weights in the formats available for your model's architecture.

<figure><img src="https://3448603185-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fcmo9mhfIjYrvFFy1U7dk%2Fuploads%2Fgit-blob-c419cf8b11bce7fc529bbf30faacc449ff875205%2Fdownload-weights.png?alt=media" alt="Download Weights button"><figcaption></figcaption></figure>

#### Method B: Python SDK

You can also use the Roboflow Python package to download weights directly to your directory:

```python
from roboflow import Roboflow

rf = Roboflow(api_key="YOUR_API_KEY")
model = rf.workspace().project("PROJECT_ID").version(1).models()[0]
model.download()  # Downloads 'weights.pt' to your local folder
```

**Note**: Roboflow does not provide technical support for model weights used outside of the Roboflow Inference ecosystem. For the best experience, we recommend using the Inference path outlined in Section 1.