Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 252 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Traffic Prediction | Los-loop | 65 | Feb 26, 2026 | ||
| Traffic Prediction | SZ-taxi | 65 | Feb 26, 2026 | ||
| Depth Estimation | ScanNet (test) | 65 | Feb 26, 2026 | ||
| Semantic Segmentation | CITYSCAPES 1/8 labeled samples 372 labels (val) | 65 | Feb 26, 2026 | ||
| Relation Extraction | GDA (test) | 65 | Feb 26, 2026 | ||
| Single Image Reflection Removal | Wild 55 images (test) | 65 | Jul 1, 2026 | ||
| Scene Text Detection | MSRA-TD500 (test) | 65 | Feb 26, 2026 | ||
| Facial Expression Recognition | AffectNet 8-way (test) | 65 | Feb 26, 2026 | ||
| Multimodal Sentiment Analysis | CMU-MOSEI standard (test) | 65 | May 19, 2026 | ||
| Image Classification | BigEarthNet (val) | 65 | Feb 26, 2026 | ||
| Node-level classification | 65 | Jun 23, 2026 | |||
| Digit Classification | MNIST -> USPS (test) | 65 | Feb 27, 2026 | ||
| Image Denoising | DND benchmark (test) | 65 | Feb 26, 2026 | ||
| Person Re-Identification | CUHK03 Labeled (767/700) | 65 | Apr 7, 2026 | ||
| 3D Object Detection | KITTI Cyclist (test) | 65 | Jun 24, 2026 | ||
| Machine Translation | WMT14 En-De newstest2014 (test) | 65 | Feb 26, 2026 | ||
| Word-level Language Modeling | WikiText-103 word-level (test) | 65 | Feb 26, 2026 | ||
| Natural Language Inference | MultiNLI matched (test) | 65 | Feb 26, 2026 | ||
| Video Interpolation | UCF-101 (test) | 65 | Feb 26, 2026 | ||
| JPEG Image Deblocking | Classic5 | 65 | Feb 26, 2026 | ||
| Individual Treatment Effect Estimation | IHDP (out-of-sample) | 65 | Jul 7, 2026 | ||
| Multi-turn Response Selection | Ubuntu Corpus | 65 | Feb 26, 2026 | ||
| Dialogue State Tracking | WOZ 2.0 (test) | 65 | Feb 26, 2026 | ||
| Reinforcement Learning | LunarLanderContinuous v2 | 65 | May 11, 2026 | ||
| Image Captioning Evaluation | Flickr8K-CF (test) | 65 | Feb 26, 2026 |