Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1762 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Semantic Textual Similarity (STS) | MTEB English 2023 (test) | 19 | Feb 26, 2026 | ||
| Face Verification | CP-LFW | 19 | Feb 26, 2026 | ||
| Instance Segmentation | OV-LVIS v1.0 (val) | 19 | Feb 26, 2026 | ||
| Code Generation | DS-1000 1.0 (test) | 19 | Feb 26, 2026 | ||
| Information Retrieval | NQ | 19 | Mar 17, 2026 | ||
| 3D Instance Segmentation | ScanNet (val) | 19 | Feb 26, 2026 | ||
| Visual Place Recognition | 17 Places | 19 | Feb 26, 2026 | ||
| Locomotion | D4RL Gym-MuJoCo walker2d-medium v2 | 19 | May 19, 2026 | ||
| Text-to-motion generation | KIT (test) | 19 | Apr 10, 2026 | ||
| Semantic Segmentation | Cityscapes 1.0 (val 1/2 split) | 19 | Mar 24, 2026 | ||
| 3D Human Pose Estimation | SceneEgo 50 (test) | 19 | Feb 26, 2026 | ||
| Adversarial Attack | ImageNet-compatible | 19 | Apr 7, 2026 | ||
| Semantic Segmentation | Pascal VOC 366 labels | 19 | Feb 26, 2026 | ||
| Molecule Design Optimization | TDC DRD3 (test) | 19 | Feb 26, 2026 | ||
| Job-Shop Scheduling Problem | JSSP 100 instances 10x10 (test) | 19 | Feb 26, 2026 | ||
| Deepfake Detection | DeepfakeTIMIT HQ | 19 | May 19, 2026 | ||
| Deepfake Detection | DeepfakeTIMIT LQ | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | YE (Yeast) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | OT (Otto) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | MI (Midas) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | JA (Jannis) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | HO (Housing) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | HI (Hill-Valley) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | HE (Heart) corrupted with Laplace noise | 19 | Feb 26, 2026 | ||
| Downstream performance prediction | GE German corrupted with Laplace noise | 19 | Feb 26, 2026 |