Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 700 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Age prediction | ADNI | 34 | May 14, 2026 | ||
| MMSE prediction | ADNI | 34 | May 14, 2026 | ||
| Multi-hop Question Answering | HotpotQA fullwiki (val) | 34 | Jul 8, 2026 | ||
| Trajectory Prediction | Hotel ETH-UCY | 34 | Jun 17, 2026 | ||
| Integer Linear Programming | MIPLIB | 34 | May 13, 2026 | ||
| UI Agent Evaluation | OSWorld | 34 | May 13, 2026 | ||
| Long-horizon web tasks | WebTailBench long-horizon tasks | 34 | May 13, 2026 | ||
| Multi-attribute steering | Multi-attribute steering S+T (test) | 34 | May 13, 2026 | ||
| Answer-position balance evaluation | MCQ | 34 | May 13, 2026 | ||
| Ensemble Committee Selection | LiveBench (test) | 34 | May 12, 2026 | ||
| Gaussian Deblurring | ImageNet 256 x 256 | 34 | Jun 30, 2026 | ||
| Health Conditions Classification | Longitudinal Wearable Sensor Waveforms | 34 | May 12, 2026 | ||
| Watermark Detection | VoxPopuli-100k Scenario T1 | 34 | Jun 16, 2026 | ||
| Instruction Following | IFEval Inverse | 34 | Jul 2, 2026 | ||
| Knowledge Graph Question Answering | WebQSP | 34 | May 11, 2026 | ||
| Structural retrieval | SCOPe-40 | 34 | May 11, 2026 | ||
| Math Reasoning | Math Reasoning Evaluation Suite (Math-500, AMC23, GSM8k, Minerva, Olympiad) | 34 | Jun 11, 2026 | ||
| Semantic Segmentation | RAW ADE20K Low-light (LOW) | 34 | May 11, 2026 | ||
| Continual Imitation Learning | LIBERO goal 1.0 | 34 | May 8, 2026 | ||
| Node Classification | CHAMELEON | 34 | May 8, 2026 | ||
| Safety Evaluation | XSTest (combined) | 34 | May 7, 2026 | ||
| Reward Estimation | Standard bus engine replacement simulation without dummy variables | 34 | May 6, 2026 | ||
| Saliency Map Evaluation | ImageNet-S50 (test) | 34 | May 4, 2026 | ||
| Attribution Faithfulness | Oxford-IIIT Pet | 34 | May 20, 2026 | ||
| Image Reasoning | WeMath | 34 | May 15, 2026 |