Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1225 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Camera Pose Estimation | ScanNet 90 frames | 24 | May 27, 2026 | ||
| Camera Pose Estimation | TUM-dynamics 90 frames | 24 | May 27, 2026 | ||
| Social Interaction Evaluation | SOTOPIA-Hard GPT-4o-as-Partner | 24 | Mar 4, 2026 | ||
| Social Interaction Evaluation | SOTOPIA GPT-4o-as-Partner | 24 | Mar 4, 2026 | ||
| Social Interaction Evaluation | SOTOPIA-Hard (Self-Play) | 24 | Mar 4, 2026 | ||
| Social Interaction Evaluation | SOTOPIA (Self-Play) | 24 | Mar 4, 2026 | ||
| Geolocation Inference | DOXBENCH | 24 | Mar 4, 2026 | ||
| Language Modeling | Mistral-7B | 24 | May 27, 2026 | ||
| Preference Prediction | PPE Preference (test) | 24 | Mar 4, 2026 | ||
| Dynamics prediction of physical systems | 2D closed chain | 24 | Mar 4, 2026 | ||
| Safety Evaluation | FigStep (test) | 24 | Mar 4, 2026 | ||
| Safety Evaluation | MM-SafetyBench v1.0 (test) | 24 | Mar 4, 2026 | ||
| Global Prediction | CIFAR-100 initial (test) | 24 | Mar 4, 2026 | ||
| Malicious client detection | Fashion-MNIST | 24 | Mar 20, 2026 | ||
| Video Question Answering | MVBench, EgoSchema, LongVideo Bench, and VideoMME | 24 | Mar 4, 2026 | ||
| Crack Segmentation | CrackMap (test) | 24 | Jun 1, 2026 | ||
| Crack Segmentation | CamCrack789 (test) | 24 | Jun 1, 2026 | ||
| Novel View Synthesis | NeRSemble | 24 | Jun 30, 2026 | ||
| Utility Evaluation | SLIMORCA (test) | 24 | Mar 4, 2026 | ||
| Safety Evaluation | ANTHROPIC HH (test) | 24 | Mar 4, 2026 | ||
| Knowledge | TruthfulQA | 24 | Jun 18, 2026 | ||
| Health-related dialogue and decision-making | HealthBench Main | 24 | May 19, 2026 | ||
| Commonsense Question Answering | CosmosQA (test) | 24 | Mar 4, 2026 | ||
| Mathematical Reasoning | MATH (test) | 24 | Mar 4, 2026 | ||
| Watermarking Detection | BOOKSUM (test) | 24 | Mar 4, 2026 |