Home / OCR / Benchmarks / coco

coco

Unknown

OCR benchmark

5
Total Results
5
Models Tested
1
Metrics
2025-12-19
Last Updated

mAP

Higher is better

Rank Model Score Source
1 co-detr-swin-l

Co-DETR with Swin-L backbone. Current SOTA range for object detection.

66 arxiv-paper
2 internimage-h

InternImage: Large-scale vision foundation model with deformable convolution.

65.4 arxiv-paper
3 dino-swin-l

DINO: DETR with Improved DeNoising Anchor Boxes.

63.3 arxiv-paper
4 yolov10-x

Real-time object detection. High efficiency.

57.4 github-readme
5 efficientdet-d7-x

EfficientDet: Scalable and Efficient Object Detection.

55.1 google-research