Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 228 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Long-form Video Understanding | EgoSchema | 69 | Jun 16, 2026 | ||
| Open Vocabulary Semantic Segmentation | PASCAL Context Context60 with background | 69 | Mar 25, 2026 | ||
| Class-Incremental Learning | CIFAR-100 B0_Inc10 | 69 | Jun 25, 2026 | ||
| Jailbreak Attack | Behaviours | 69 | Feb 26, 2026 | ||
| Person Re-Identification | CUHK03 NP (test) | 69 | Feb 26, 2026 | ||
| Chinese Multitask Language Understanding | CMMLU | 69 | Jun 30, 2026 | ||
| Text-to-Image Retrieval | RSICD (test) | 69 | Jul 8, 2026 | ||
| OOD Detection | CIFAR100 Dfar | 69 | Apr 15, 2026 | ||
| Visual-Inertial Odometry | EuRoC (All sequences) | 69 | Jun 23, 2026 | ||
| Large Multimodal Model Evaluation | MM-Vet | 69 | Apr 24, 2026 | ||
| Language Modeling | WikiText (test) | 69 | Jul 9, 2026 | ||
| Text-to-Image Generation | COCO 2014 (val) | 69 | Jul 7, 2026 | ||
| Few-shot Image Classification | 11 datasets average CLIP-based (ImageNet, Caltech101, OxfordPets, StanfordCars, Flowers102, Food101, FGVCAircraft, SUN397, DTD, EuroSAT, UCF101) | 69 | Mar 26, 2026 | ||
| Continual Learning | CIFAR100 (test) | 69 | May 6, 2026 | ||
| Object Tracking | COESOT (test) | 69 | May 8, 2026 | ||
| Classification | AID (test) | 69 | Jun 9, 2026 | ||
| Time-series forecasting | Illness | 69 | Mar 11, 2026 | ||
| Low-Light Image Enhancement | LOL v1 | 69 | Mar 19, 2026 | ||
| Out-of-Distribution Detection | ImageNet-1k (ID) with 4 OOD datasets (iNaturalist, SUN, Places, Textures) | 69 | May 27, 2026 | ||
| Stereo Image Super-Resolution | KITTI 2012 (test) | 69 | Jul 7, 2026 | ||
| Change Detection | S2Looking (test) | 69 | Feb 27, 2026 | ||
| Lung segmentation | Shenzhen dataset (test) | 69 | Feb 26, 2026 | ||
| Image Generation | AFHQ v2 | 69 | Jul 3, 2026 | ||
| Arithmetic reasoning | ASDiv | 69 | Jun 4, 2026 | ||
| Few-shot classification | MiniImageNet 5-way 10-shot (test) | 69 | Feb 26, 2026 |