Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2072 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Partial Label Learning | CIFAR-10 | 17 | Mar 4, 2026 | ||
| Image Watermarking | DiffusionDB | 17 | Feb 26, 2026 | ||
| OOD Detection | OpenOOD CIFAR100 as ID v1.5 (test) | 17 | May 15, 2026 | ||
| Semi-supervised Video Object Segmentation | YTVOS 2019 (val) | 17 | May 19, 2026 | ||
| Semi-supervised Video Object Segmentation | DAVIS 17 (test-dev) | 17 | Feb 26, 2026 | ||
| Multi-channel Image Classification | So2Sat Full channels (test) | 17 | Mar 17, 2026 | ||
| Image Restoration | DIV2K (val) | 17 | Feb 26, 2026 | ||
| Long-tailed Image Classification | iNat (val test) | 17 | Feb 26, 2026 | ||
| Generalized Category Discovery | Semantic Shift Benchmark (SSB) | 17 | Feb 26, 2026 | ||
| Action Recognition | Kinetics standard (meta-test) | 17 | Feb 26, 2026 | ||
| Sub-activity Classification | MOMA Sub-activities (meta-test) | 17 | Feb 26, 2026 | ||
| Emotion Recognition | CREMA-D 6-class | 17 | Feb 26, 2026 | ||
| Part Segmentation | PACO-Part | 17 | Mar 31, 2026 | ||
| Automated Audio Captioning | AudioCaps (evaluation) | 17 | Jul 7, 2026 | ||
| Long Video Question Answering | EgoSchema (full set) | 17 | Feb 26, 2026 | ||
| Card Games | Point24 | 17 | Feb 26, 2026 | ||
| Offline Model-Based Optimization | D'Kitty | 17 | Feb 26, 2026 | ||
| Offline Model-Based Optimization | ChEMBL | 17 | Feb 26, 2026 | ||
| Offline Model-Based Optimization | UTR | 17 | Feb 26, 2026 | ||
| Offline Model-Based Optimization | GFP | 17 | Feb 26, 2026 | ||
| Offline Model-Based Optimization | TF-Bind-8 | 17 | Feb 26, 2026 | ||
| Seq2Seq generation | QQP | 17 | Feb 26, 2026 | ||
| Backdoor Defense | CIFAR-10 LC DenseNet-161 (test) | 17 | Feb 26, 2026 | ||
| Backdoor Defense | CIFAR-10 SSBA DenseNet-161 (test) | 17 | Feb 26, 2026 | ||
| Backdoor Defense | CIFAR-10 LC v1 (test) | 17 | Feb 26, 2026 |