Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1511 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Semantic Segmentation | SYNTHIA source to CityScapes, BDD, and Mapillary targets 16 shared categories (val) | 21 | Feb 26, 2026 | ||
| Image Classification | CIFAR-100 semantic noise (test) | 21 | Feb 26, 2026 | ||
| Image Classification | ImageNet 1k (val) | 21 | Feb 26, 2026 | ||
| Fine-grained Image Classification | OxfordPets (test) | 21 | Feb 26, 2026 | ||
| 3D Dense Captioning | ScanRefer | 21 | Mar 31, 2026 | ||
| Scene Text Visual Question Answering | ST-VQA (test) | 21 | Feb 26, 2026 | ||
| Hand Manipulation | Adroit door-expert | 21 | Feb 26, 2026 | ||
| Normal Estimation | SceneNN (test) | 21 | Feb 26, 2026 | ||
| Action Recognition | Charades-Ego first-person (test) | 21 | Feb 26, 2026 | ||
| Image Classification | Patching Task Suite Average | 21 | Feb 26, 2026 | ||
| Task-incremental Learning | CIFAR10 (test) | 21 | Feb 26, 2026 | ||
| Question Generation | SQuAD | 21 | Feb 26, 2026 | ||
| Depth Estimation | Taskonomy (test) | 21 | Feb 26, 2026 | ||
| Point-of-no-return temporal localization | Ego4D v1 (test) | 21 | Feb 26, 2026 | ||
| Multimodal Named Entity Recognition | Twitter-15 (test) | 21 | Feb 26, 2026 | ||
| Long-tailed Image Classification | ImageNet-LT 1.0 (test) | 21 | Feb 26, 2026 | ||
| Long-tailed Image Recognition | CIFAR100 LT IF 100 (test) | 21 | Feb 26, 2026 | ||
| Predicate Classification | VG-1800 (test) | 21 | Feb 26, 2026 | ||
| Image Classification | CIFAR-100-LT (Imbalance 20) | 21 | Feb 26, 2026 | ||
| Image Classification | CIFAR-100-LT Imbalance 200 | 21 | Feb 26, 2026 | ||
| Natural Language Queries | Ego4D NLQ (test) | 21 | Feb 26, 2026 | ||
| Image Classification | Stylized-ImageNet (test) | 21 | Jul 7, 2026 | ||
| Continual Image Classification | ImageNet-1000 10 steps | 21 | Feb 26, 2026 | ||
| Super-Resolution | Sun80 (test) | 21 | Feb 26, 2026 | ||
| Object Counting | FSCD-LVIS (test) | 21 | Feb 27, 2026 |