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 |
|---|---|---|---|---|---|
| 3D Human Pose Estimation | CMU Panoptic (test) | 47 | Jun 8, 2026 | ||
| Anomaly Localization | MVTec AD 1.0 (test) | 47 | Mar 25, 2026 | ||
| Multi-organ segmentation | Synapse | 47 | May 12, 2026 | ||
| Low-light Image Enhancement | VV | 47 | Apr 8, 2026 | ||
| Object Classification | ScanObjectNN PB_T50_RS (test) | 47 | Feb 26, 2026 | ||
| Offline Reinforcement Learning |
|---|
| D4RL antmaze-large (diverse) |
| 47 |
| May 15, 2026 |
| Offline Reinforcement Learning | D4RL antmaze-large (play) | 47 | Jun 11, 2026 |
|---|
| Semantic Segmentation | COCO-Stuff-10K (test) | 47 | Feb 26, 2026 |
|---|
| Referring Expression Comprehension | ReferItGame (test) | 47 | Feb 26, 2026 |
|---|
| Blind Face Restoration | CelebA (test) | 47 | Mar 4, 2026 |
|---|
| OOD detection | Textures (OOD) | 47 | Mar 4, 2026 |
|---|
| Human Mesh Recovery | Human3.6M | 47 | Feb 26, 2026 |
|---|
| Video Super-Resolution | Vimeo-90K-T BI degradation (test) | 47 | Feb 26, 2026 |
|---|
| Semi-supervised Domain Adaptation | Office-Home 3-shot | 47 | Feb 26, 2026 |
|---|
| Dialog State Tracking | MultiWOZ 2.0 (test) | 47 | Feb 26, 2026 |
|---|
| Abductive Natural Language Inference | aNLI (leaderboard) | 47 | Feb 26, 2026 |
|---|
| Semantic Image Synthesis | ADE20K (val) | 47 | Feb 26, 2026 |
|---|
| LiDAR-based Panoptic Segmentation | SemanticKITTI (test) | 47 | Feb 26, 2026 |
|---|
| Anomaly Detection | MNIST one-class classification | 47 | Feb 26, 2026 |
|---|
| Video Summarization | TVSum (test) | 47 | Feb 26, 2026 |
|---|
| Findings Generation | IU-Xray (test) | 47 | Feb 26, 2026 |
|---|
| Person Search | CUHK-PEDES (test) | 47 | Feb 26, 2026 |
|---|
| Abstractive Summarization | Multi-News | 47 | Feb 26, 2026 |
|---|
| Semantic Segmentation | BDD (val) | 47 | Feb 26, 2026 |
|---|
| Real Image Denoising | SIDD (val) | 47 | Apr 3, 2026 |
|---|