| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Anomaly Detection | ATLAS | AUC-ROC81.5 | 20 | |
| Interactive Image Segmentation | ATLAS T1 | Dice Coefficient86 | 18 | |
| Knowledge grounding | ATLAS CC-QKV | ACC@1100 | 16 | |
| Knowledge grounding | ATLAS-Pes2o-QKV | ACC@1100 | 16 | |
| Surgical video segmentation | ATLAS-120k (test) | AP64 | 15 | |
| Unsupervised Brain MRI Anomaly Detection | ATLAS | Dice26.67 | 14 | |
| Brain Lesion Detection | ATLAS T1 | DICE25.5 | 12 | |
| Anomaly Localization | ATLAS | AP0.33 | 12 | |
| Trace Reconstruction | ATLAS (D3) | Recall80.8 | 10 | |
| Liver and Tumor Segmentation | ATLAS MRI (test) | Dice Score (Liver)96.5 | 10 | |
| Lesion Segmentation | ATLAS (test) | DSC53.49 | 9 | |
| Voxel-level anomaly segmentation | ATLAS T1w R2.0 | AUROC96.2 | 8 | |
| Patient-level anomaly detection | ATLAS T1w R2.0 | AUROC100 | 8 | |
| McKean-Vlasov system prediction | Atlas 100D (test) | W1 Distance17.345 | 8 | |
| Medical Image Segmentation | ATLAS | Dice Score63.8 | 8 | |
| Ischemic stroke lesion segmentation | ATLAS v2.0 (test) | AUPRC19.96 | 8 | |
| Medical Image Segmentation | ATLAS 2 | Dice Score63 | 7 | |
| Brain Stroke Segmentation | ATLAS (val) | Dice48.67 | 7 | |
| Story Generation | Atlas 100 pages | Story Count89 | 6 | |
| Story Generation | Atlas 20 pages | Story Score78.66 | 6 | |
| Story Generation | Atlas 2 pages | Story Count92.33 | 6 | |
| Semantic Segmentation | ATLAS | DSC87.5 | 6 | |
| Interleaved Sequence Demixing | ATLAS | Accuracy42.3 | 6 | |
| Visual Question Answering | ATLAS 2.0 | Modality Accuracy97.7 | 6 | |
| Report Generation | ATLAS 2.0 | ROUGE33.69 | 6 |