Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1504 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Commonsense Question Answering | Abductive NLI (aNLI) (val) | 21 | Feb 26, 2026 | ||
| Customized Text-to-Image Generation | DreamBench (test) | 21 | Apr 13, 2026 | ||
| RGBT Tracking | VTUAV | 21 | Feb 26, 2026 | ||
| Gaussian Deblurring | CelebA-HQ 1K (test) | 21 | Feb 26, 2026 | ||
| Sign Language Production | How2Sign (test) | 21 | Jun 4, 2026 | ||
| Zero-shot Classification | EuroSAT | 21 | May 19, 2026 | ||
| Underwater Salient Object Detection | USOD10k 1.0 (test) | 21 | Feb 26, 2026 | ||
| Unsupervised Person Re-identification | Market1501 | 21 | Feb 26, 2026 | ||
| Class-agnostic 3D instance segmentation | ScanNet++ | 21 | Jun 15, 2026 | ||
| Object Recognition | OpenImages v7 (val) | 21 | Feb 26, 2026 | ||
| Object Recognition | CC3M (test) | 21 | Feb 26, 2026 | ||
| Vision-and-Language Navigation | IR2R (val-unseen) | 21 | Mar 31, 2026 | ||
| Vision-and-Language Navigation | IR2R (val-seen) | 21 | Mar 31, 2026 | ||
| CT Image Reconstruction | AAPM OOD (test) | 21 | Feb 26, 2026 | ||
| CT Image Reconstruction | Cropped OOD circle (test) | 21 | Feb 26, 2026 | ||
| Novel View Synthesis | X3D Multiple scenes (Average reported) | 21 | Feb 26, 2026 | ||
| Semantic Segmentation | Cityscapes original (1/2 (1488 labeled images)) | 21 | Jul 7, 2026 | ||
| Instance Segmentation | COCO | 21 | Jun 16, 2026 | ||
| LiDAR Scene Flow Estimation | Waymo Open Dataset 1.0 (val) | 21 | Feb 26, 2026 | ||
| Scientific data visualization | MatPlotBench 1.0 (test) | 21 | Feb 26, 2026 | ||
| Competitive Programming | Codeforces Rating | 21 | Feb 26, 2026 | ||
| Multiple-choice question answering | MMLU (test) | 21 | Jul 2, 2026 | ||
| Medical Report Generation | IU X-Ray | 21 | Feb 26, 2026 | ||
| Adversarial Defense | ModelNet40 | 21 | Feb 26, 2026 | ||
| Visual Reasoning | NLVR2 (dev) | 21 | Apr 21, 2026 |