# YOLO26-N

> YOLO26-N: one-stage object detector from the yolo26 family. Planned for LibreYOLO (not yet packaged).

- Family: yolo26
- Parameters: 2.4M
- GFLOPs: 5.4
- Default input size: 640px
- Detection: anchor-free · nms-free
- Architecture: one-stage · CSPDarknet backbone
- Weight license: non-permissive
- Paper-reported mAP@50-95: 40.9%

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 40.1 | 10.4 | 96.5 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 40.1 | 21.6 | 46.4 | 119 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 40.1 | 60.2 | 16.6 | 11 |
| Raspberry Pi 5 | ncnn FP32 | 40.1 | 12.9 | 77.6 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 40.1 | 7.1 | 140.6 | - |
| Raspberry Pi 5 | PyTorch FP32 | 40.1 | 5.0 | 198.1 | - |


Source: https://www.visionanalysis.org/model/yolo26n. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).