Ultralytics YOLO
Open-source YOLO models (YOLO26, YOLO11) for detection, segmentation, pose and tracking, plus the Ultralytics Platform for training and deployment
Overview
Ultralytics makes the most widely used open-source YOLO models and the Python package that trains and runs them. YOLO26, announced in January 2026, and earlier YOLO11 and YOLOv8 cover object detection, instance segmentation, classification, pose estimation, oriented bounding boxes and tracking; YOLO27 is listed as "Coming Soon" with no release date. Code and models are AGPL-3.0, so closed-source or SaaS use needs a paid Enterprise License. The hosted Ultralytics Platform adds dataset management, cloud training and deployment. Founder and CEO Glenn Jocher leads the company, which has offices in Frederick (Maryland), London and Madrid and raised a $30M Series A led by Elephant in September 2025. For a hosted dataset and labeling platform see Roboflow.
The Verdict
Who Should Use Ultralytics YOLO?
Best For
- Engineers who need fast, real-time object detection on edge or cloud
- Open-source projects that can live with AGPL-3.0
- Teams that want train, export and deploy in one Python package
Not Ideal For
- Closed-source products that will not buy an Enterprise License
- Open-vocabulary or language-driven vision tasks
What's Great
- One package for detect, segment, pose, OBB and tracking
- Very active project with a large community
- Free Platform tier and low-cost Pro seats
Watch Out For
- AGPL-3.0 also covers models you fine-tune, a frequent complaint
- Enterprise License price is not public
- Licensing rules confuse small teams
Pricing
View all features & details
Key Features
- Models: YOLO26, YOLO11, YOLOv8
- Detection, segmentation, classification, pose, OBB, tracking
- Python API and CLI
- Export to edge and mobile formats
Use Cases
- Manufacturing inspection
- Retail and traffic analytics
- Robotics and drones
- Medical and agriculture imaging
How It Compares
| Feature | Ultralytics YOLO | Roboflow | Landing AI |
|---|---|---|---|
| Main job | Open-source vision models | Datasets, labeling, hosted training | Document and visual AI |
| License | AGPL-3.0 or Enterprise | Commercial SaaS | Commercial SaaS |
| Self-host | Yes | Partly | See vendor |
| Entry paid plan | $29/seat/mo | See vendor | See vendor |
Sources: Ultralytics license · Ultralytics about · GitHub · Pricing
Is this your tool? Add a badge