skip to content
gigarouter gigarouter
models / object detection · coming soon

YOLO11

Ultralytics/YOLO11

published Oct 2024 · updated Jun 2026

YOLO11 is a state-of-the-art object detection model from Ultralytics that is fast, accurate, and easy to use.

status
coming soon
API providers
0
downloads / mo
9.8K
license
agpl-3.0

specs

TaskObject Detection
ArchitectureYOLO11
Parameters2.6M (nano) to 56.9M (x-large)
LicenseAGPL-3.0 (Enterprise license available for commercial use)

about this model

Ultralytics YOLO11 is a real-time object detection model that identifies and localizes objects within images and video frames. Released on September 30, 2024, it is the latest iteration in the YOLO series, designed for high-speed inference with strong accuracy across 80 object classes from the COCO dataset.

Key Strengths

  • Parameter efficiency: The YOLO11m variant achieves higher mean average precision (mAP) on COCO while using 22% fewer parameters than its predecessor YOLOv8m.
  • Multiple model sizes: Five variants (n, s, m, l, x) allow trade-offs between speed and accuracy, from the lightweight YOLO11n (2.6M parameters) to the high-accuracy YOLO11x (56.9M parameters).
  • Multi-task support: Beyond detection, YOLO11 supports instance segmentation, pose estimation, image classification, oriented object detection (OBB), and tracking.

Benchmark Performance (COCO val2017)

Model mAP 50-95 Speed (CPU ONNX, ms) Speed (T4 TensorRT10, ms) Parameters (M)
YOLO11n 39.5 56.1 1.5 2.6
YOLO11s 47.0 90.0 2.5 9.4
YOLO11m 51.5 183.2 4.7 20.1
YOLO11l 53.4 238.6 6.2 25.3
YOLO11x 54.7 462.8 11.3 56.9

Benchmark speeds are averaged over COCO val images using an Amazon EC2 P4d instance. CPU speeds measured with ONNX export; GPU speeds measured with TensorRT export. All models operate at 640x640 pixel input size.

Capabilities

  • Detection on 80 COCO object classes.
  • Oriented object detection (OBB) for rotated bounding boxes.
  • Instance segmentation and pose estimation (separate model variants).
  • Image classification pretrained on ImageNet.

Gigarouter hosts YOLO11 as a managed, OpenAI-compatible API. No local installation or model loading is required — send inference requests directly to the endpoint.

best for

FAQ

What is the input format for the YOLO11 detection API?

The API accepts image data (base64-encoded or URL) and returns bounding boxes, class labels, and confidence scores for detected objects.

How does YOLO11 compare to YOLOv8 in terms of efficiency?

YOLO11m achieves higher mAP on COCO while using 22% fewer parameters than YOLOv8m.

What is the license for YOLO11?

YOLO11 is released under the AGPL-3.0 license. An enterprise license is available for commercial use from Ultralytics.

How can I call YOLO11 via the gigarouter API?

Use the gigarouter OpenAI-compatible endpoint with your API key to send images and receive detection results.

What are the model size variants available for YOLO11 detection?

YOLO11 offers five detection variants: nano (2.6M params), small (9.4M), medium (20.1M), large (25.3M), and x-large (56.9M).

not yet live

We're benchmarking and onboarding YOLO11 as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.

related object detection models

compare all →