At matched compute on an RTX 5070 Ti, RT-DETR is the more accurate family: it averages 1.3 mAP points higher and wins 6 of 7 nearest-parameter pairs. YOLOX answers on speed, running 30.3% faster on average in PyTorch. Pick by which axis you optimize for.
RT-DETR fields 7 measured variants; YOLOX fields 6. Both run the same COCO val2017 protocol at 640 px on an RTX 5070 Ti, so the two ladders align by parameter count.
| Model | Family | Params (M) | mAP@50-95 | FPS |
|---|---|---|---|---|
| YOLOX-Nano | YOLOX | 0.9 | 2876.0 | 53.6 |
| YOLOX-Tiny | YOLOX | 5.1 | 3547.0 | 61.6 |
| YOLOX-S | YOLOX | 9.0 | 4427.0 | 50.0 |
| RT-DETR-R18 | RT-DETR | 20.2 | 4979.0 | 47.2 |
| YOLOX-M | YOLOX | 25.3 | 5169.0 | 48.2 |
| RT-DETR-R34 | RT-DETR | 31.4 | 5223.0 | 40.2 |
| RT-DETR-L | RT-DETR | 32.9 | 5577.0 | 27.7 |
| RT-DETR-R50m | RT-DETR | 36.6 | 5382.0 | 35.3 |
| RT-DETR-R50 | RT-DETR | 42.9 | 5588.0 | 29.8 |
| YOLOX-L | YOLOX | 54.2 | 5542.0 | 42.8 |
| RT-DETR-X | RT-DETR | 67.4 | 5794.0 | 21.8 |
| RT-DETR-R101 | RT-DETR | 76.6 | 5677.0 | 23.6 |
| YOLOX-X | YOLOX | 99.1 | 5627.0 | 39.6 |
Accuracy at matched compute
mAP is shown in percent. Pairing by nearest parameter count, RT-DETR wins 6 of 7 pairs and averages 1.3 mAP points higher. At the top end RT-DETR-R101 reaches 56.8 mAP against YOLOX-X at 56.3. The single pair YOLOX takes is narrow.
Speed
YOLOX is the faster family: 30.3% faster on average in PyTorch FP32. Its smallest variant, YOLOX-Nano, runs 53.6 FPS at 0.91M parameters, above any RT-DETR variant's throughput here.
- RT-DETR
- Apache-2.0 (permissive)
- YOLOX
- Apache-2.0 (permissive)
- Commercial use
- Both families ship under Apache-2.0
Which family to pick
Pick RT-DETR when accuracy per parameter leads: it wins nearly every matched pair. Pick YOLOX when throughput is the constraint, or when you need the smallest models, down to 0.91M parameters. Both ship under Apache-2.0, so licensing does not force the call. See the per-variant specs on the RT-DETR and YOLOX 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.