Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1559 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Question Answering | 2WikiMulti | 20 | May 19, 2026 | ||
| Table Question Answering | STQA L | 20 | May 19, 2026 | ||
| Table Question Answering | STQA-N | 20 | May 19, 2026 | ||
| Infrared Small Target Detection | DAUB | 20 | May 19, 2026 | ||
| Rhetorical Role Labeling | DEEPRHOLE (test) | 20 | May 19, 2026 | ||
| Rhetorical Role Labeling | SCOTUSSteps (test) | 20 | May 19, 2026 | ||
| Rhetorical Role Labeling | SCOTUSRF (test) | 20 | May 19, 2026 | ||
| Visual Question Answering | RSME-Bench | 20 | May 19, 2026 | ||
| Visual Question Answering | CHOICE | 20 | May 19, 2026 | ||
| Medical Image Segmentation | Abd-CT (Setting 1) | 20 | Jun 16, 2026 | ||
| Language Modeling | ClimbMix FineWeb-Edu 1M-token (val) | 20 | May 19, 2026 | ||
| Text-to-Image Generation | OCR | 20 | Jun 30, 2026 | ||
| Machine Learning Efficiency | Adult, Default, Shoppers, Magic, Beijing, and News | 20 | May 19, 2026 | ||
| Object Segmentation | HRSOD | 20 | May 19, 2026 | ||
| Object Segmentation | ThinObject5K (test) | 20 | May 19, 2026 | ||
| Object Segmentation | COIFT | 20 | May 19, 2026 | ||
| Object Segmentation | DIS5K VD | 20 | May 19, 2026 | ||
| Math Reasoning | AMC23, AMC24, AIME24, AIME25, MATH500 | 20 | May 19, 2026 | ||
| Multiclass classification | Lycos 2017 | 20 | May 19, 2026 | ||
| Malicious Tool Call Detection | TRUST-BENCH reconstructed comparison set curated 1,970-episode (test) | 20 | May 19, 2026 | ||
| Medical Image Segmentation | MRI (47 anatomical and pathological targets) SAMRI Performance Evaluation (test) | 20 | May 19, 2026 | ||
| Long-context Language Understanding | LongBench | 20 | May 19, 2026 | ||
| Function Calling | BFCL Memory | 20 | May 19, 2026 | ||
| Stochastic route planning | SOP (Stochastic Orienteering Problem) N = 500 nodes | 20 | May 19, 2026 | ||
| Zero-shot Classification | 20-task CLIP evaluation suite (test) | 20 | May 19, 2026 |