Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1555 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Adaptive Retrieval Privacy Attack Success | 500 nearest-neighbor and random prefix pairs mid-network layer | 20 | May 26, 2026 | ||
| Commonsense and Logical Reasoning | Overall CSQA, StrategyQA, LogiQA | 20 | May 26, 2026 | ||
| Logical Reasoning | LogiQA | 20 | May 26, 2026 | ||
| Commonsense Reasoning | StrategyQA | 20 | May 26, 2026 | ||
| Commonsense Reasoning | CSQA | 20 | May 26, 2026 | ||
| Mathematical Reasoning | MathQA | 20 | May 26, 2026 | ||
| Mathematical Reasoning | SVAMP | 20 | May 26, 2026 | ||
| Named Entity Recognition | i2b2 2014 (test) | 20 | May 26, 2026 | ||
| Edited Image Captioning | MMEdit E-IC | 20 | May 25, 2026 | ||
| Edited Visual Question Answering | MMEdit E-VQA | 20 | May 25, 2026 | ||
| High-dimensional optimization | LIMO | 20 | May 25, 2026 | ||
| High-dimensional optimization | Lasso-Hard | 20 | May 25, 2026 | ||
| Open-ended Writing | WritingBench | 20 | Jun 1, 2026 | ||
| Weight-space classification | Model Jungle | 20 | May 25, 2026 | ||
| Question Answering | HotpotQA τ = 96k | 20 | May 25, 2026 | ||
| Long-context Performance Evaluation | LoCoMo tau = 96k (test) | 20 | May 25, 2026 | ||
| Image Classification | CIFAR-100-LTN Asymmetric Noise IF=10, NR=40% (test) | 20 | May 25, 2026 | ||
| Image Classification | CIFAR-100-LTN Symmetric Noise IF=10, NR=40% (test) | 20 | May 25, 2026 | ||
| Binary Classification | STL10 | 20 | May 25, 2026 | ||
| Out-of-Distribution Detection | CIFAR-10 (In-distribution) vs SVHN (OOD) (test) | 20 | Jul 9, 2026 | ||
| Classification | CIFAR100C (test) | 20 | May 25, 2026 | ||
| Multi-class classification | CIFAR100 LT imbalance factor 0.1 (test) | 20 | May 25, 2026 | ||
| Multi-turn Conversation Evaluation | IMDb | 20 | May 25, 2026 | ||
| Multi-turn Conversation Evaluation | EduFeedback alternate | 20 | May 25, 2026 | ||
| Multi-turn Conversation Evaluation | UltraFeedback | 20 | May 25, 2026 |