Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2083 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Selective Visual Question Answering | Mixed ID/OOD 50% VQA v2 / 50% AdVQA (test) | 17 | Feb 26, 2026 | ||
| Selective Visual Question Answering | Mixed ID/OOD 66.7% VQA v2 / 33.3% AdVQA (test) | 17 | Feb 26, 2026 | ||
| Discriminative tasks | BigBench 13 tasks (val) | 17 | Feb 26, 2026 | ||
| Video Question Answering | TGIF-QA Action original (test) | 17 | Feb 26, 2026 | ||
| Joint Optic Cup and Optic Disc Segmentation | BASE3 | 17 | Feb 26, 2026 | ||
| Joint Optic Cup and Optic Disc Segmentation | BASE2 | 17 | Feb 26, 2026 | ||
| Joint Optic Cup and Optic Disc Segmentation | BASE1 | 17 | Feb 26, 2026 | ||
| Interaction Attribution | Identity Rule language | 17 | Feb 26, 2026 | ||
| Classification | Australian (UCI) (test) | 17 | May 18, 2026 | ||
| Classification | Ionosphere (UCI) (test) | 17 | May 18, 2026 | ||
| Intent Classification | HWU64 | 17 | Mar 11, 2026 | ||
| Intent Classification | CLINC150 | 17 | Mar 11, 2026 | ||
| 3D MRI reconstruction | HR 3D MRI 380x294x138x24 | 17 | Feb 26, 2026 | ||
| Machine Translation | WMT news Average 2010-2016 (test) | 17 | Feb 26, 2026 | ||
| Speech Translation | Callhome En-Es (test) | 17 | Feb 26, 2026 | ||
| Binary classification | AdvGLUE (test) | 17 | Feb 26, 2026 | ||
| Full-Reference Image Quality Assessment | KADID | 17 | Feb 26, 2026 | ||
| Object Detection | M6Doc (test) | 17 | Feb 26, 2026 | ||
| Relation Extraction | RETACRED (test) | 17 | Feb 26, 2026 | ||
| Referring Expression Segmentation | G-Ref UMD (test) | 17 | Feb 26, 2026 | ||
| Math Word Problem Solution Generation | GSM8K (test) | 17 | Feb 26, 2026 | ||
| Compositional Reasoning | ARO | 17 | Jun 4, 2026 | ||
| Grammatical Error Correction | NLPCC word-level (test) | 17 | Feb 26, 2026 | ||
| Commonsense Fact Verification | CREAK (test) | 17 | Jun 4, 2026 | ||
| Source Separation | WSJ0-2mix (test) | 17 | Feb 26, 2026 |