Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1771 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Targeted Adversarial Attack | MNIST | 19 | Feb 26, 2026 | ||
| Fine-grained Image Classification | CUB 5-way | 19 | Feb 26, 2026 | ||
| Depth Estimation | 7-Scenes (test) | 19 | Feb 26, 2026 | ||
| Case Marker Extraction | Silver Standard | 19 | Feb 26, 2026 | ||
| Semantic Segmentation | Cityscapes 1/2 partition protocol (val) | 19 | Feb 26, 2026 | ||
| HOI Detection | HICO-DET Unseen Combination | 19 | Feb 27, 2026 | ||
| 6D Pose Estimation | Occlusion dataset BOP challenge (test) | 19 | Feb 26, 2026 | ||
| Text-to-Image Synthesis | CelebA-HQ (test) | 19 | Feb 26, 2026 | ||
| Visual Grounding | ReferCOCOg UMD (test-u) | 19 | Feb 26, 2026 | ||
| Visual Grounding | ReferCOCO v1 (val) | 19 | Feb 26, 2026 | ||
| Action Recognition | PA-HMDB | 19 | Mar 30, 2026 | ||
| Stereo Matching | Middlebury quarter resolution (test) | 19 | Feb 26, 2026 | ||
| Semantic Segmentation | DADA-seg 1.0 (test) | 19 | Feb 26, 2026 | ||
| Action Classification | Kinetics 400 2017 (test) | 19 | Feb 26, 2026 | ||
| Semantic Segmentation | DensePASS (val) | 19 | Feb 26, 2026 | ||
| Portrait Matting | PPM-100 (test) | 19 | Feb 26, 2026 | ||
| Image Classification | C-MNIST (test) | 19 | Jun 4, 2026 | ||
| Image Classification | ImageSquawk 128x128 (val) | 19 | May 14, 2026 | ||
| Image Classification | ImageMeow 128x128 (val) | 19 | May 14, 2026 | ||
| Object Detection | MS-COCO Standard 10% 2017 (val) | 19 | Feb 26, 2026 | ||
| Sarcasm explanation generation | WITS (test) | 19 | Feb 26, 2026 | ||
| Unpaired Image-to-Image Translation | Zebra -> Horse CycleGAN (test) | 19 | Feb 26, 2026 | ||
| Unpaired Image-to-Image Translation | CycleGAN Horse -> Zebra (test) | 19 | Feb 26, 2026 | ||
| Table Question Answering | WIKISQL WEAK (dev) | 19 | Feb 26, 2026 | ||
| Landmark Detection | Head X-ray dataset (test) | 19 | Feb 26, 2026 |