Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 718 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Machine Reading Comprehension | CMRC 2018 (dev) | 34 | Feb 26, 2026 | ||
| Classification | RT | 34 | Mar 5, 2026 | ||
| Image Captioning | Conceptual Captions (test) | 34 | Feb 26, 2026 | ||
| Speech Recognition | WSJ nov92 (test) | 34 | Feb 26, 2026 | ||
| Named Entity Recognition | GENIA (test) | 34 | Feb 26, 2026 | ||
| Image Retrieval | Tokyo 24/7 (test) | 34 | Feb 26, 2026 | ||
| Visual Localization | Extended CMU Seasons Urban | 34 | Feb 26, 2026 | ||
| Question Answering | SQuAD 2.0 (test) | 34 | Feb 26, 2026 | ||
| Question Answering | HotpotQA distractor setting (test) | 34 | Jun 12, 2026 | ||
| Steering angle prediction | Udacity (test) | 34 | Apr 7, 2026 | ||
| Out-of-distribution Detection | SVHN in-distribution LSUN out-of-distribution (test) | 34 | Feb 26, 2026 | ||
| Human Body Part Segmentation | LSP (test) | 34 | Feb 26, 2026 | ||
| Named Entity Recognition | ONTONOTES (test) | 34 | Feb 26, 2026 | ||
| Image Dehazing | RESIDE SOTS | 34 | Feb 26, 2026 | ||
| Image Classification | WikiArt | 34 | Apr 8, 2026 | ||
| Addressee and Response Selection | Ubuntu Multiparty Conversation Corpus v1 (test) | 34 | Feb 26, 2026 | ||
| Addressee and Response Selection | Ubuntu Multiparty Conversation Corpus v1 (dev) | 34 | Feb 26, 2026 | ||
| Movie Fill-in-the-Blank | LSMDC 2016 (test) | 34 | Feb 26, 2026 | ||
| Zero-Shot Recognition | AwA (test) | 34 | Feb 26, 2026 | ||
| Image Retrieval | Paris 106k (Par106k) | 34 | Feb 26, 2026 | ||
| Word Alignment | Romanian-English (Ro-En) (test) | 34 | Feb 26, 2026 | ||
| Zero-shot Classification | CUB 2011 (test) | 34 | Feb 26, 2026 | ||
| Paraphrase Detection | MSRP | 34 | Feb 26, 2026 | ||
| Permutation Invariant Image Classification | MNIST (test) | 34 | Feb 26, 2026 | ||
| Molecule description generation | ChEBI-20 (test) | 34 | Feb 26, 2026 |