Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1509 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Semantic Image Editing | ADE20K-Room | 21 | Feb 26, 2026 | ||
| Semantic Image Editing | Cityscapes | 21 | Feb 26, 2026 | ||
| Question Answering | SQuAD v2.0 (val) | 21 | Feb 26, 2026 | ||
| Instance Segmentation | ADE20K (val) | 21 | Feb 27, 2026 | ||
| Novel View Synthesis | DTU 9-view (test) | 21 | Feb 26, 2026 | ||
| Video Compression | Standard Video Compression Suite UVG, MCL-JCV, HEVC B/C/D/E/RGB | 21 | Feb 26, 2026 | ||
| 3D Object Detection | KITTI car category (val) | 21 | Feb 26, 2026 | ||
| Video Recognition | UCF101 v1 (test) | 21 | Feb 26, 2026 | ||
| Gaussian Denoising | Set12 | 21 | Feb 26, 2026 | ||
| Barycenter Computation | Location-scatter population Uniform P0 = Uniform([-sqrt(3), +sqrt(3)]^D) (test) | 21 | Feb 26, 2026 | ||
| Barycenter Computation | Location-scatter population (Gaussian P0 = N(0, ID)) (test) | 21 | Feb 26, 2026 | ||
| Activity Detection | Charades (val) | 21 | Feb 26, 2026 | ||
| Scene Classification | NWPU 10/90 split | 21 | Feb 26, 2026 | ||
| Long-form Transcription | CORAAL | 21 | Feb 26, 2026 | ||
| Speech Recognition | Multilingual LibriSpeech (MLS) (test) | 21 | Mar 13, 2026 | ||
| Image Classification | ImageNet ID | 21 | Feb 26, 2026 | ||
| Spatiotemporal Action Localization | AVA 2.2 | 21 | Feb 26, 2026 | ||
| Teachable Object Recognition | ORBIT CLU-VE 1.0 (test) | 21 | Feb 26, 2026 | ||
| Teachable Object Recognition | ORBIT CLE-VE 1.0 (test) | 21 | Feb 26, 2026 | ||
| Long-text classification | Hyperpartisan news detection (test) | 21 | Feb 26, 2026 | ||
| Video Object Segmentation | Long-time Video dataset (val) | 21 | Feb 26, 2026 | ||
| Simulability | ML-QE (test) | 21 | Feb 26, 2026 | ||
| Navigation | Obstacle World (test) | 21 | Feb 26, 2026 | ||
| Regression | Wine | 21 | May 19, 2026 | ||
| Regression | Power | 21 | Jul 9, 2026 |