# YOLO26-S

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

- Family: yolo26
- Parameters: 9.5M
- GFLOPs: 20.7
- Default input size: 640px
- Detection: anchor-free · nms-free
- Architecture: one-stage · CSPDarknet backbone
- Weight license: non-permissive
- Paper-reported mAP@50-95: 48.6%

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 47.4 | 4.0 | 247.5 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 47.4 | 20.1 | 49.9 | 353 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 47.4 | 41.7 | 24.0 | 11 |
| Raspberry Pi 5 | ncnn FP32 | 47.4 | 5.7 | 176.9 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 47.4 | 2.8 | 360.4 | - |
| Raspberry Pi 5 | PyTorch FP32 | 47.4 | 2.3 | 441.8 | - |


Source: https://www.visionanalysis.org/model/yolo26s. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).