Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 985 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Phrase grounding | Flickr30k Entities (test) | 28 | Feb 26, 2026 | ||
| Question Answering | SQuAD-Open | 28 | Feb 26, 2026 | ||
| Aspect term-polarity pair extraction | SemEval Laptop (SL) (test) | 28 | Feb 26, 2026 | ||
| Question Answering | MetaQA 2-hop | 28 | Mar 17, 2026 | ||
| Speech Recognition | Wall Street Journal open vocabulary (dev93) | 28 | Feb 26, 2026 | ||
| Action Recognition | HMDB51 (all splits) | 28 | Feb 26, 2026 | ||
| Sentence Representation Evaluation | SentEval (test) | 28 | Feb 26, 2026 | ||
| Spectral Clustering | MNIST | 28 | Feb 26, 2026 | ||
| Image Classification | pixel-by-pixel MNIST (test) | 28 | Feb 27, 2026 | ||
| Semantic Role Labeling | CoNLL English Brown 2009 (test) | 28 | Feb 26, 2026 | ||
| Image Inpainting | Paris StreetView (test) | 28 | Feb 26, 2026 | ||
| Chunking | Chunk (test) | 28 | Feb 26, 2026 | ||
| Multi-Person Pose Tracking | PoseTrack Challenge Leaderboard 2017 (test) | 28 | Feb 26, 2026 | ||
| Link Prediction | Kinship (test) | 28 | Jun 1, 2026 | ||
| Digit Classification | SVHN -> MNIST | 28 | Feb 27, 2026 | ||
| Solving reaction equations | Reaction equations | 28 | Feb 26, 2026 | ||
| Classification | TCGA-NSCLC subtyping | 28 | Feb 26, 2026 | ||
| Face Verification | MegaFace Challenge 1 | 28 | Feb 26, 2026 | ||
| Facial Landmark Detection | 300-W Public Common subset inter-pupil normalization (test) | 28 | Feb 26, 2026 | ||
| Facial Landmark Detection | 300-W public (test) | 28 | Feb 26, 2026 | ||
| Action Recognition | UT-Kinect | 28 | Feb 26, 2026 | ||
| Question Answering | SearchQA (dev) | 28 | Feb 26, 2026 | ||
| Semantic Segmentation | NYU-depth v2 | 28 | Apr 15, 2026 | ||
| Unsupervised Domain Adaptation | MNIST -> USPS (test) | 28 | Feb 26, 2026 | ||
| Video Deinterlacing | Video Deinterlacing Sequences (test) | 28 | Feb 26, 2026 |