Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2758 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| WiFi-based Gesture Recognition | Real-world collected dataset In-Domain | 14 | Mar 24, 2026 | ||
| Deepfake Detection | DF40 | 14 | Apr 10, 2026 | ||
| Video Editing | DAVIS (first 33 frames) | 14 | Feb 26, 2026 | ||
| Object Retexture | UHRSD (test) | 14 | Feb 26, 2026 | ||
| Video Object Retexturing | Pexels video dataset (test) | 14 | Feb 26, 2026 | ||
| Trajectory Planning | Unified Evaluation Settings Autonomous Driving (test) | 14 | Feb 26, 2026 | ||
| 3D Object Classification | ScanObjectNN OBJ_ONLY v1.0 | 14 | Feb 26, 2026 | ||
| 3D Object Classification | ScanObjectNN OBJ_BG v1.0 | 14 | Feb 26, 2026 | ||
| Long Video Question Answering | Video-MME w/o subtitles | 14 | Feb 26, 2026 | ||
| Image Classification | Caltech | 14 | Mar 4, 2026 | ||
| Glacial Lake Segmentation | Glacial Lake (Multiple Lakes) | 14 | Feb 26, 2026 | ||
| Glacial Lake Segmentation | Glacial Lake Single Lake | 14 | Feb 26, 2026 | ||
| Human Avatar Reconstruction | Our constructed database (Novel view) | 14 | Feb 26, 2026 | ||
| Video Question Answering | MovieChat Global Breakpoint | 14 | Feb 26, 2026 | ||
| Anomaly Segmentation | Real-IAD | 14 | Jul 2, 2026 | ||
| HOI Video Generation | HOIGen-1M 1.0 (test) | 14 | Feb 26, 2026 | ||
| Slide-level classification | MLH1-UCEC (test) | 14 | Feb 26, 2026 | ||
| Slide-level classification | P53-UCEC (test) | 14 | Feb 26, 2026 | ||
| Scene Flow Estimation | VoD Radar evaluation (val) | 14 | May 19, 2026 | ||
| Video Compression | 1080p videos | 14 | Feb 26, 2026 | ||
| Video Compression | HEVC Class C | 14 | Feb 26, 2026 | ||
| Multi-view Camera Pose Estimation | BlendedMVS | 14 | Feb 26, 2026 | ||
| Multi-view Camera Pose Estimation | ScanNet++ | 14 | Feb 26, 2026 | ||
| Open-vocabulary attribute personalization | Abstract Concepts | 14 | Feb 26, 2026 | ||
| Open-vocabulary attribute personalization | Concrete Objects | 14 | Feb 26, 2026 |