#2433·yolov3

New YOLO models live in ultralytics/ultralytics

Author: glenn-jocherCreated Jun 15, 2026Updated Jun 15, 2026
Labelsdocumentationdetect

YOLOv3 remains available here for legacy projects, historical reproducibility, and users who specifically need this repository.

For all new Ultralytics models, current documentation, active development, and production workflows, use the main repository:

https://github.com/ultralytics/ultralytics

Use the main repo for new projects

The ultralytics/ultralytics repository is the canonical home for the current Ultralytics YOLO family, including:

  • YOLO26, the latest stable recommended model generation
  • YOLO11 and YOLOv8 support
  • Detection, segmentation, semantic segmentation, classification, pose, OBB, and tracking
  • Current training, validation, prediction, export, and deployment workflows
  • Latest docs, examples, bug fixes, and releases

Install the current package

bash
pip install -U ultralytics

Run YOLO26 from the current package

bash
# Predict
yolo predict model=yolo26n.pt source=https://ultralytics.com/images/bus.jpg imgsz=640

# Train
yolo train data=coco8.yaml model=yolo26n.pt epochs=10 imgsz=640

# Export
yolo export model=yolo26n.pt format=onnx imgsz=640

Python example

python
from ultralytics import YOLO

model = YOLO("yolo26n.pt")
results = model("https://ultralytics.com/images/bus.jpg")

Current docs and resources

When should I still use this YOLOv3 repo?

Use this repository when you intentionally need the YOLOv3 codebase, historical YOLOv3 behavior, or old project compatibility.

For new models, new features, current docs, and active maintenance, start in ultralytics/ultralytics.