Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1505 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Jailbreak Attack | Advbench Vicuna-33B Guard 100 prompts Original | 21 | Feb 26, 2026 | ||
| Jailbreak Attack | Advbench Llama2-70B Guard 100 prompts Original | 21 | Feb 26, 2026 | ||
| Entity Recognition | BioNLP2004, MIT Movie, MIT Restaurant, and OntoNotes 5.0 (Average) few-shot (test) | 21 | Feb 26, 2026 | ||
| Image Classification | Visda 2021 | 21 | Jul 2, 2026 | ||
| Image Matching | ScanNet 1500 | 21 | May 6, 2026 | ||
| WSI-level Search | Private-Liver Internal (test) | 21 | Feb 26, 2026 | ||
| WSI-level Search | Private-Breast Internal (test) | 21 | Feb 26, 2026 | ||
| Hand-Object Pose Estimation | DexYCB | 21 | Jul 3, 2026 | ||
| Face Forgery Detection | FaceForensics++ (DeepFakes) | 21 | Feb 26, 2026 | ||
| Long-term Time Series Forecasting | ETTm1 to ETTh2 | 21 | Feb 26, 2026 | ||
| CTR prediction | MovieLens (test) | 21 | Feb 26, 2026 | ||
| Conversational Search Retrieval | TopiOCQA (test) | 21 | Feb 26, 2026 | ||
| Depth Image Classification | NYU-D | 21 | May 19, 2026 | ||
| Multi-person motion prediction | ExPI unseen action | 21 | Feb 26, 2026 | ||
| Segmentation | TNMIX | 21 | Feb 26, 2026 | ||
| Segmentation | REFUGE2 Optic-Cup | 21 | Feb 26, 2026 | ||
| Segmentation | REFUGE2 Optic-Disc | 21 | Feb 26, 2026 | ||
| Image-to-Text Retrieval | Flickr30K English 1K (test) | 21 | Feb 26, 2026 | ||
| Image Classification | DTD (Novel) | 21 | May 12, 2026 | ||
| Action Anticipation | EGTEA Gaze+ | 21 | Jul 7, 2026 | ||
| Audio-Visual Segmentation | AVSBench S4 (test) | 21 | Mar 25, 2026 | ||
| Semantic Object Identification | COST (val) | 21 | Feb 26, 2026 | ||
| Graph Clustering | Chameleon | 21 | Jun 30, 2026 | ||
| Semantic Segmentation | Cityscapes, BDD100K, and Mapillary Aggregate (test) | 21 | Feb 26, 2026 | ||
| Fixel-based analysis | HCP 8 subjects | 21 | Feb 26, 2026 |