Neither family wins outright on an RTX 5070 Ti. YOLOv9 leads at the small end and runs 22.8% faster on average in PyTorch; RF-DETR leads at the large end, where RF-DETR-L reaches 58.6 mAP against YOLOv9-C at 57.1. Across 4 nearest-parameter pairs the accuracy split is 2 to 2, with YOLOv9 averaging 1.5 mAP points higher.
RF-DETR fields 4 measured variants; YOLOv9 fields 4. All run the same COCO val2017 protocol on an RTX 5070 Ti. RF-DETR's variants sit between 30.47M and 33.93M parameters, so each pairs with YOLOv9-C (25.5M) by count.
| Model | Family | Params (M) | mAP@50-95 | FPS |
|---|---|---|---|---|
| YOLOv9-T | YOLOv9 | 2.0 | 41.8 | 31.9 |
| YOLOv9-S | YOLOv9 | 7.2 | 50.4 | 30.8 |
| YOLOv9-M | YOLOv9 | 20.1 | 56.1 | 37.1 |
| YOLOv9-C | YOLOv9 | 25.5 | 57.1 | 42.3 |
| RF-DETR-N | RF-DETR | 30.5 | 51.4 | 39.5 |
| RF-DETR-S | RF-DETR | 32.1 | 55.1 | 35.8 |
| RF-DETR-M | RF-DETR | 33.7 | 57.3 | 30.3 |
| RF-DETR-L | RF-DETR | 33.9 | 58.6 | 25.1 |
Accuracy at matched compute
mAP is shown in percent. Against YOLOv9-C at 57.1 mAP, RF-DETR-N trails at 51.4 and RF-DETR-S at 55.1. RF-DETR-M leads by 0.2 points at 57.3 and RF-DETR-L by 1.5 at 58.6. That is a 2 to 2 split, with YOLOv9 ahead by 1.5 points on average. At the small end YOLOv9-T reaches 41.8 mAP at 2.04M parameters, a size class RF-DETR does not field.
Speed
YOLOv9 is faster at matched compute: 22.8% faster on average in PyTorch FP32. YOLOv9-C runs 42.3 FPS; RF-DETR runs 39.5 FPS at RF-DETR-N and 25.1 FPS at RF-DETR-L.
Where the frontier crosses
The accuracy lead changes hands with model size. At the small end YOLOv9 leads: its parameter-efficient variants sit above RF-DETR's smallest, RF-DETR-N, which needs 30.47M parameters to reach 51.4 mAP. At the large end RF-DETR leads, with RF-DETR-L at 58.6 mAP. There is no single winner; the choice follows your size class.
- RF-DETR
- Apache-2.0 (permissive)
- YOLOv9
- MIT (permissive)
- Commercial use
- Both families ship under permissive licenses
Which family to pick
Pick YOLOv9 for small and mid-size deployments and for throughput: it fields models from 2.04M parameters and is faster at matched size. Pick RF-DETR-L when peak accuracy matters more than frame rate: it reaches 58.6 mAP, 1.5 points above YOLOv9-C, at 25.1 vs 42.3 FPS. Both ship under permissive licenses. See the per-variant specs on the RF-DETR and YOLOv9 model pages.
Every number on this page comes from the verified dataset: same 500-image COCO val2017 slice, conf 0.001, IoU 0.6, max 300 detections, pycocotools mAP, identical protocol across all hardware and runtimes. The full protocol is on the methodology page. To rerun this comparison with your own filters, open compare. Accuracy is measured on LibreYOLO retrained checkpoints; other weight sources can yield different values.
