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 |
|---|---|---|---|---|---|
| Image Classification | ImageNet-1K IPC 50 | 30 | Apr 21, 2026 | ||
| Function Calling | BFCL Individual Tools per Problem | 30 | Feb 26, 2026 | ||
| Mortality Prediction | PhysioNet Challenge 2012 | 30 | Jun 30, 2026 | ||
| Question Answering | PopQA | 30 | Mar 25, 2026 | ||
| Watermark Detection | GSM8K | 30 | Feb 26, 2026 | ||
| PII-Real |
| 30 |
| Feb 26, 2026 |
| Query-Related PII Detection | PII-multi | 30 | Feb 26, 2026 |
|---|
| Query-Related PII Detection | PII-Real (test) | 30 | Feb 26, 2026 |
|---|
| Query-Unrelated PII Masking | PII-single and PII-multi | 30 | Feb 26, 2026 |
|---|
| Query-Related PII Detection | PII-single 1.0 (test) | 30 | Feb 26, 2026 |
|---|
| Node Classification | Cora HOMO. (test) | 30 | Feb 26, 2026 |
|---|
| Node Classification | Cora HET. (test) | 30 | Feb 26, 2026 |
|---|
| Multi-task Language Understanding and Reasoning | OpenCompass SIQA, GSM8K, WiC, HumanEval, MMLU, CSQA | 30 | Feb 26, 2026 |
|---|
| Multi-Task Reinforcement Learning (LTL Instruction Following) | Warehouse Finite Horizon | 30 | Feb 26, 2026 |
|---|
| Needle-in-a-Haystack Retrieval | Needle-in-a-Haystack 32K context (test) | 30 | Feb 26, 2026 |
|---|
| Needle-in-a-Haystack Retrieval | Needle-in-a-Haystack 8K context (test) | 30 | Feb 26, 2026 |
|---|
| Image Reconstruction | ImageNet 50K 256x256 (val) | 30 | Jun 23, 2026 |
|---|
| Semantic Alignment | Church | 30 | Feb 26, 2026 |
|---|
| Explanation Generation | Hindi News v1.0 (test) | 30 | Feb 26, 2026 |
|---|
| Table Question Answering | HiTab | 30 | Feb 26, 2026 |
|---|
| Clustering | Statlog | 30 | May 12, 2026 |
|---|
| Time Series Forecasting | AusRain | 30 | Feb 26, 2026 |
|---|
| Surgical Smoke Removal | DesmokeData (test) | 30 | Mar 30, 2026 |
|---|
| Spatio-Temporal Video Grounding | HC-STVG v1 (test) | 30 | Jun 16, 2026 |
|---|
| Multi-hop Question Answering | MuSiQue (val) | 30 | Jul 8, 2026 |
|---|