Loading the SOTA2 catalog…
SOTA2 Research · benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
| Task | Dataset | Domain | Trend | Results | Last Update |
|---|---|---|---|---|---|
| PDE Solving | Helmholtz equation | 37 | Jun 16, 2026 | ||
| Domain Generalization | DomainNet DomainBed (test) | 37 | Feb 27, 2026 | ||
| Machine Translation | WMT En-De 2019 (test) | 37 | Feb 26, 2026 | ||
| Liver Segmentation | LiTS | 37 | Jun 16, 2026 | ||
| Multilingual text-centric visual question answering | MTVQA | 37 | Apr 15, 2026 | ||
| Fine-grained classification |
| Describable Textures Dataset (DTD) |
| 37 |
| May 18, 2026 |
| Point Cloud Completion | PCN | 37 | Mar 24, 2026 |
|---|
| Generation | MLLMU-Bench (Forget Set) | 37 | Jun 2, 2026 |
|---|
| Fine grained classification | Oxford-IIIT Pets | 37 | May 13, 2026 |
|---|
| Phase recognition | Cholec80 (test) | 37 | Jul 1, 2026 |
|---|
| Video Instance Segmentation | YouTube-VIS 2022 (val) | 37 | Jun 19, 2026 |
|---|
| Conversational Performance | REDIAL (test) | 37 | Feb 26, 2026 |
|---|
| Long document summarization | BookSum (test) | 37 | Feb 26, 2026 |
|---|
| Audio-Visual Segmentation | AVSBench MS3 v1 (test) | 37 | Feb 26, 2026 |
|---|
| Visual Question Answering | ScienceVQA | 37 | Jun 12, 2026 |
|---|
| Segmentation | Cityscapes | 37 | Mar 20, 2026 |
|---|
| Inductive Link Prediction | FB15k-237 inductive (test) | 37 | Feb 26, 2026 |
|---|
| Hyperspectral Image Classification | Pavia University | 37 | Jun 4, 2026 |
|---|
| Image Classification | ImageNet-100 original (test) | 36 | Jul 9, 2026 |
|---|
| PDE Solving | Burgers' nu=0.001 1D (test) | 36 | Jul 9, 2026 |
|---|
| Visual Question Answering | IR-VQA 9,720-sample (test) | 36 | Jul 8, 2026 |
|---|
| 3D Dense Captioning | ScanRefer | 36 | Jul 8, 2026 |
|---|
| Source Code Watermark Extraction Robustness | CodeNet D152 (evaluation) | 36 | Jul 8, 2026 |
|---|
| Source Code Watermark Extraction Robustness | CodeNet D88 (evaluation) | 36 | Jul 8, 2026 |
|---|
| Source Code Watermark Extraction Robustness | CodeNet D56 (evaluation) | 36 | Jul 8, 2026 |
|---|