Models Supported by Ultralytics#
Welcome to Ultralytics' model documentation! We offer support for a wide range of models, each tailored to specific tasks like object detection, instance segmentation, semantic segmentation, depth estimation, image classification, pose estimation, and multi-object tracking. If you're interested in contributing your model architecture to Ultralytics, check out our Contributing Guide.
Featured Models#
Start with YOLO26 for a new project: it is the latest Ultralytics release and the only one covering all seven tasks. YOLO11 is the mature alternative, with pretrained checkpoints for each of its five tasks. Pick a specialized family only when you need promptable segmentation (SAM 3), open-vocabulary detection (YOLOE, YOLO-World), or a transformer detector (RT-DETR).
The table lists every documented model with the tasks it covers, which of the train, val, predict and export modes Ultralytics supports for it, and when to choose it. Track is not listed separately: it runs on top of predict for Detect, Segment, Pose and OBB models, while SAM 2 and SAM 3 track through their own video predictors. Benchmark is not listed either, because it wraps export and val across formats rather than adding support of its own.
| Model | Tasks | Modes | Choose it for |
|---|---|---|---|
| YOLO26 🚀 NEW | Detect, Segment, Semantic, Depth, Classify, Pose, OBB | Train, Val, Predict, Export | New projects: end-to-end NMS-free inference, edge-optimized deployment, and the widest task coverage |
| YOLO12 | Detect, Segment, Classify, Pose, OBB | Train, Val, Predict, Export | Benchmarking the attention-centric community release; pretrained weights cover detection only |
| YOLO11 | Detect, Segment, Classify, Pose, OBB | Train, Val, Predict, Export | Stable production workloads, with pretrained checkpoints for every task it supports |
| YOLOv10 | Detect | Train, Val, Predict, Export | NMS-free detection research from Tsinghua University |
| YOLOv9 | Detect, Segment | Train, Val, Predict, Export | Programmable Gradient Information (PGI), implemented on the Ultralytics YOLOv5 codebase |
| YOLOv8 | Detect, Segment, Classify, Pose, OBB | Train, Val, Predict, Export | Existing YOLOv8 pipelines and established third-party integrations |
| YOLOv7 | Detect | Predict | Running an upstream-trained YOLOv7 export: the package predicts from a compatible ONNX or TensorRT model, never from the native checkpoint |
| YOLOv6 | Detect | Train, Val, Predict, Export | Training Meituan's architecture from its YAML; Ultralytics hosts no YOLOv6 .pt checkpoints |
| YOLOv5 | Detect | Train, Val, Predict, Export | Legacy Ultralytics projects; current releases load the updated YOLOv5u checkpoints |
| YOLOv4 | None | None | Architecture reference only: Alexey Bochkovskiy's Darknet-native model is not supported by the package |
| YOLOv3 | Detect | Train, Val, Predict, Export | Legacy projects on Joseph Redmon's original architecture; current releases load the updated YOLOv3u checkpoints |
| SAM 3 🚀 NEW | Segment | Predict | Meta's promptable concept segmentation in images and video, from text or image exemplars; sam3.pt requires access approval on Hugging Face |
| SAM 2 | Segment | Predict | Meta's promptable segmentation, tracking objects across video frames |
| SAM | Segment | Predict | Meta's original promptable segmentation, including auto-annotation |
| MobileSAM | Segment | Predict | Promptable segmentation on mobile and other resource-constrained devices (Kyung Hee University) |
| FastSAM | Segment | Val, Predict, Export | CNN-based promptable segmentation when SAM latency is the bottleneck (Chinese Academy of Sciences) |
| YOLO-NAS | Detect | Val, Predict, Export | Deci's NAS-optimized detectors, kept for inference and export; Deci no longer maintains them after the NVIDIA acquisition |
| RT-DETR | Detect | Train, Val, Predict, Export | Baidu's real-time DETR detector: a convolutional backbone with a hybrid transformer encoder |
| YOLO-World | Detect | Train, Val, Predict, Export | Open-vocabulary detection from text prompts (Tencent AI Lab); export requires the -worldv2 checkpoints |
| YOLOE | Detect, Segment | Train, Val, Predict, Export | Open-vocabulary detection and segmentation with text, visual, or prompt-free inference |
Watch: Run Ultralytics YOLO models in just a few lines of code.
Getting Started: Usage Examples#
This example provides simple YOLO training and inference examples. For full documentation on these and other modes see the Predict, Train, Val and Export docs pages.
Note the below example spotlights YOLO26 Detect models for object detection. For additional supported tasks see the Segment, Semantic, Depth, Classify, Pose and OBB docs.
PyTorch pretrained *.pt models as well as configuration *.yaml files can be passed to the YOLO(), SAM(), NAS() and RTDETR() classes to create a model instance in Python:
from ultralytics import YOLO
# Load a COCO-pretrained YOLO26n model
model = YOLO("yolo26n.pt")
# Display model information (optional)
model.info()
# Train the model on the COCO8 example dataset for 100 epochs
results = model.train(data="coco8.yaml", epochs=100, imgsz=640)
# Run inference with the YOLO26n model on the 'bus.jpg' image
results = model("path/to/bus.jpg")Contributing New Models#
Interested in contributing your model to Ultralytics? Great! We're always open to expanding our model portfolio.
-
Fork the Repository: Start by forking the Ultralytics GitHub repository.
-
Clone Your Fork: Clone your fork to your local machine and create a new branch to work on.
-
Implement Your Model: Add your model following the coding standards and guidelines provided in our Contributing Guide.
-
Test Thoroughly: Make sure to test your model rigorously, both in isolation and as part of the pipeline.
-
Create a Pull Request: Once you're satisfied with your model, create a pull request to the main repository for review.
-
Code Review & Merging: After review, if your model meets our criteria, it will be merged into the main repository.
For detailed steps, consult our Contributing Guide.
FAQ#
What is the latest Ultralytics YOLO model?#
The latest Ultralytics YOLO model is YOLO26, released in January 2026. YOLO26 features end-to-end NMS-free inference, optimized edge deployment, and supports detection, instance segmentation, semantic segmentation, depth estimation, classification, pose estimation, and OBB plus open-vocabulary versions. For stable production workloads, both YOLO26 and YOLO11 are recommended choices.
How can I train a YOLO model on custom data?#
Training a YOLO model on custom data can be easily accomplished using Ultralytics' libraries. Here's a quick example:
from ultralytics import YOLO
# Load a YOLO model
model = YOLO("yolo26n.pt") # or any other YOLO model
# Train the model on custom dataset
results = model.train(data="custom_data.yaml", epochs=100, imgsz=640)For more detailed instructions, visit the Train documentation page.
Which YOLO versions are supported by Ultralytics?#
Ultralytics natively supports YOLOv3, YOLOv5, YOLOv6, YOLOv8, YOLOv9, YOLOv10, YOLO11, YOLO12, and YOLO26, along with the SAM family (SAM 3, SAM 2, SAM, MobileSAM and FastSAM), YOLO-NAS, RT-DETR, YOLO-World, and YOLOE. The package publishes no weights or YAMLs for YOLOv4 or YOLOv7: YOLOv4 is documented as an architecture reference only, while YOLOv7 runs as an exported ONNX or TensorRT model. See Featured Models for the tasks and modes available for each.
Why should I use Ultralytics Platform for machine learning projects?#
Ultralytics Platform provides a no-code, end-to-end platform for training, deploying, and managing YOLO models. It simplifies complex workflows, enabling users to focus on model performance and application. It also offers cloud training capabilities, comprehensive dataset management, and user-friendly interfaces for both beginners and experienced developers.
What types of tasks can Ultralytics YOLO models perform?#
Ultralytics YOLO models are versatile and can perform tasks including object detection, instance segmentation, semantic segmentation, depth estimation, classification, pose estimation, and oriented object detection (OBB). The latest model, YOLO26, supports all seven tasks plus open-vocabulary detection. For details on specific tasks, refer to the Task pages.