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 |
|---|---|---|---|---|---|
| Retrosternal area localization | Cardiac CT images 200 experiments per cell 512 x 512 pixels | 36 | May 22, 2026 | ||
| Time-Series Classification | SKODA | 36 | Jun 23, 2026 | ||
| Time-Series Classification | HARTH | 36 | Jun 23, 2026 | ||
| Collaborative coefficient estimation | Synthetic multiple-response linear models (test) | 36 | May 21, 2026 | ||
| Image Classification | ImageWoof ImageNet-1K (test) |
| 36 |
| May 21, 2026 |
| Driver subsidy control | Brazilian cities dataset City C (test) | 36 | May 20, 2026 |
|---|
| Driver subsidy control | Brazilian cities dataset City B (test) | 36 | May 20, 2026 |
|---|
| Driver subsidy control | Brazilian cities dataset City A (test) | 36 | May 20, 2026 |
|---|
| Visual Planning for Assistance | Visual Planning for Assistance | 36 | May 20, 2026 |
|---|
| Face Identification | Privacy-Commons | 36 | May 20, 2026 |
|---|
| Knowledge Evaluation | C-Eval (test) | 36 | May 19, 2026 |
|---|
| General AI Assistant Task | GAIA | 36 | Jun 9, 2026 |
|---|
| Backdoor defense | OGB-arxiv | 36 | May 19, 2026 |
|---|
| Backdoor defense | Flickr | 36 | May 19, 2026 |
|---|
| Backdoor defense | Pubmed | 36 | May 19, 2026 |
|---|
| Backdoor defense | Cora | 36 | May 19, 2026 |
|---|
| Drug-Drug Interaction prediction | ZhangDDI | 36 | May 19, 2026 |
|---|
| Multi-label classification | WM Nicol | 36 | May 19, 2026 |
|---|
| Trojan Attack | InjecAgent | 36 | May 19, 2026 |
|---|
| Prompt Injection | MATH | 36 | May 19, 2026 |
|---|
| Prompt Injection | CSQA | 36 | May 19, 2026 |
|---|
| Data Attribution | CIFAR-10 | 36 | Jun 5, 2026 |
|---|
| Multi-target Visual Active Search | DOTA (test) | 36 | May 18, 2026 |
|---|
| Knowledge Retention | WMDP retain | 36 | May 15, 2026 |
|---|
| Adversarial Attack | HarmBench (HB) and JailbreakBench (JBB) (held-out prompts) | 36 | May 15, 2026 |
|---|