Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1205 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Class-agnostic 3D instance segmentation | ScanNet200 | 24 | Jun 15, 2026 | ||
| Uncertainty Estimation | OBQA | 24 | Apr 13, 2026 | ||
| Uncertainty Estimation | RACE Llama-3.1-8B and Gemma-2-9B backbones (test) | 24 | Apr 13, 2026 | ||
| State Variable Inference | Chua’s circuit system | 24 | Apr 13, 2026 | ||
| Image Manipulation Localization | Autosplice | 24 | May 28, 2026 | ||
| Image Manipulation Localization | CocoGlide | 24 | May 28, 2026 | ||
| Video Temporal Grounding | Charades-STA | 24 | May 6, 2026 | ||
| Prompt Injection Defense | SEP | 24 | May 28, 2026 | ||
| Prompt Injection Defense | OPI (Open-Prompt-Injection) | 24 | May 28, 2026 | ||
| Hyperspectral Image Denoising | Washington DC Mall HSI Gaussian + Stripe noise | 24 | Apr 10, 2026 | ||
| HSI Denoising | Salinas HSI | 24 | Apr 10, 2026 | ||
| OOD Detection | GastroVision | 24 | Apr 10, 2026 | ||
| Deepfake Detection | Protocol 2 Hybrid, FR, FS, EFS v1 (test) | 24 | Apr 10, 2026 | ||
| Deepfake Detection | Protocol 1 (FF++, DFDCP, DFD, CDF2) v1 (test) | 24 | Apr 10, 2026 | ||
| Transfer Entropy Estimation | 70-dimensional linear stacking linear Gaussian system, Y → X, T=10000 | 24 | Apr 10, 2026 | ||
| Transfer Entropy Estimation | 70-dimensional linear stacking linear Gaussian system, X → Y, T=10000 | 24 | Apr 10, 2026 | ||
| Classification | Heart | 24 | May 26, 2026 | ||
| Time-series generation | MetroPT3 | 24 | Apr 10, 2026 | ||
| Time-series generation | Sine | 24 | Apr 10, 2026 | ||
| Time-series generation | ETT | 24 | Apr 10, 2026 | ||
| Time-series generation | Electric Motor (EM) | 24 | Apr 10, 2026 | ||
| Synthetic time series generation | MetroPT3 | 24 | Apr 10, 2026 | ||
| Synthetic time series generation | Sine | 24 | Apr 10, 2026 | ||
| Synthetic time series generation | ETT | 24 | Apr 10, 2026 | ||
| Synthetic time series generation | ECG | 24 | Apr 10, 2026 |