Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 546 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Watermarking | LFQA | 40 | Apr 8, 2026 | ||
| Watermarking | C4 | 40 | Apr 8, 2026 | ||
| General Reasoning | Overall | 40 | Apr 10, 2026 | ||
| Jailbreak Attack Safety Evaluation | GPT-as-a-Judge | 40 | Apr 7, 2026 | ||
| Jailbreak Attack Safety Evaluation | CARES adversarial mental-health targets 2025 | 40 | Apr 7, 2026 | ||
| Language Understanding | MMLU | 40 | May 29, 2026 | ||
| Image Classification | Caltech101 | 40 | May 26, 2026 | ||
| Sequential Recommendation | Yelp (test) | 40 | Jun 4, 2026 | ||
| Image Captioning Evaluation | ICBench Long Caption | 40 | Apr 7, 2026 | ||
| Summarization Evaluation | SummEval | 40 | Apr 7, 2026 | ||
| Unsupervised Image Hashing | NUS-WIDE | 40 | Apr 7, 2026 | ||
| Unsupervised Image Hashing | COCO | 40 | Apr 7, 2026 | ||
| VRU Separability | H-City (test) | 40 | Apr 7, 2026 | ||
| Membership Inference Attack | 4 datasets aggregate | 40 | Apr 6, 2026 | ||
| Treatment Effect Estimation | IHDP disconnected regime (m1 = 0) (test) | 40 | Apr 6, 2026 | ||
| Medical Visual Question Answering | Pooled Medical VQA Closed-ended | 40 | Apr 6, 2026 | ||
| Zero-shot Reasoning | ARC-e, PIQA, OpenbookQA, Winogrande, HellaSwag, MathQA | 40 | Jun 19, 2026 | ||
| Untargeted identity retrieval attack | VGGFace2 | 40 | Apr 2, 2026 | ||
| Membership Inference Attack | MIMIR HackerNews | 40 | Apr 1, 2026 | ||
| Membership Inference Attack | MIMIR DM Mathematics | 40 | Apr 1, 2026 | ||
| Membership Inference Attack | MIMIR Arxiv | 40 | Apr 1, 2026 | ||
| Membership Inference Attack | MIMIR Pile CC | 40 | Apr 1, 2026 | ||
| Membership Inference Attack | MIMIR PubMed Central | 40 | Apr 1, 2026 | ||
| Membership Inference Attack | MIMIR Wikipedia | 40 | Apr 1, 2026 | ||
| Microscopic weight completion | CD | 40 | Apr 1, 2026 |