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 Reconstruction | Mip-NeRF 360 (test) | 44 | Jun 8, 2026 | ||
| Skeleton-based Action Recognition | NW-UCLA | 44 | Feb 26, 2026 | ||
| SAR Object Detection | SSDD | 44 | Jun 1, 2026 | ||
| Malicious Fine-tuning Defense | BeaverTails (test) | 44 | Feb 26, 2026 | ||
| Visual Object Tracking | LaSoT | 44 | Feb 26, 2026 | ||
| Anatomy-conditioned Image Retrieval |
| MIMIC-IR official (test) |
| 44 |
| Feb 27, 2026 |
| Needle-in-a-Haystack | Needle-in-a-Haystack | 44 | Feb 26, 2026 |
|---|
| Polyp Segmentation | CVC-300 (Unseen) | 44 | Apr 21, 2026 |
|---|
| Multimodal Emotion Recognition in Conversation | IEMOCAP 6-class (test) | 44 | Mar 11, 2026 |
|---|
| Reasoning | SIQA | 44 | Feb 26, 2026 |
|---|
| Survival Prediction | TCGA-STAD (5-fold cross-validation) | 44 | Feb 26, 2026 |
|---|
| Factual Knowledge | Bool Q | 44 | Mar 24, 2026 |
|---|
| Fine-grained Perception | HR-Bench 4K | 43 | Jul 8, 2026 |
|---|
| Skin Lesion Segmentation | ISIC 2017 | 43 | Jul 8, 2026 |
|---|
| Few-Shot Domain Incremental Learning | DomainNet | 43 | Jun 30, 2026 |
|---|
| Training Efficiency Profiling | Qwen3 family S=512 (train profiling) | 43 | Jun 23, 2026 |
|---|
| System Prompt Corruption | System Prompt Corruption (SPC) | 43 | Jun 23, 2026 |
|---|
| Question Answering | Question Answering | 43 | Jun 23, 2026 |
|---|
| Agent Tool Selection | Agent Tool Selection (ATS) | 43 | Jun 23, 2026 |
|---|
| Science Reasoning | GPQA | 43 | Jul 2, 2026 |
|---|
| Radiology Report Meta-Evaluation | ReEvalMed | 43 | Jun 18, 2026 |
|---|
| General Video Understanding | TempCompass | 43 | Jul 2, 2026 |
|---|
| Image Classification | AG-MNIST (test) | 43 | Jun 16, 2026 |
|---|
| Multiple Choice Question Answering | Arc-Easy v1 (test) | 43 | Jun 12, 2026 |
|---|
| Node Classification | Cora | 43 | Jul 7, 2026 |
|---|