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 |
|---|---|---|---|---|---|
| Text-to-image | MMA-Diffusion (test) | 39 | Jun 5, 2026 | ||
| Photorealistic Style Transfer | Novel content-style pairs | 39 | Jun 2, 2026 | ||
| Mathematical Reasoning | AMC23 | 39 | Jun 30, 2026 | ||
| Safety Evaluation | Harmbench | 39 | Jun 2, 2026 | ||
| Science Question Answering | ScienceQA |
| 39 |
| Jun 24, 2026 |
| Rescheduling | EGLESE scenarios 57-112 1.0 (Part-B) | 39 | Jun 1, 2026 |
|---|
| Video Understanding | MVBench | 39 | Jun 11, 2026 |
|---|
| Automated Red-Teaming of LLM-Based Search Agents | SAFESEARCH | 39 | May 29, 2026 |
|---|
| Transductive Dynamic Link Prediction | UN Vote | 39 | May 29, 2026 |
|---|
| Transductive Dynamic Link Prediction | Myket | 39 | May 29, 2026 |
|---|
| Transductive Dynamic Link Prediction | MOOC | 39 | May 29, 2026 |
|---|
| Event-Driven Financial Forecasting | S&P500 latest (test) | 39 | May 28, 2026 |
|---|
| Table Reasoning | TABLEBENCH | 39 | May 29, 2026 |
|---|
| Question Answering | ARC-e | 39 | Jun 26, 2026 |
|---|
| Single-turn child safety evaluation | KIDBench single-turn | 39 | May 26, 2026 |
|---|
| Semi-supervised image classification | CUB200 (10% train) | 39 | May 26, 2026 |
|---|
| Language-conditioned Exploration | Matterport3D Habitat Simulation | 39 | May 25, 2026 |
|---|
| Time Series Forecasting | PEMS03 | 39 | May 22, 2026 |
|---|
| Image Classification Calibration | PACS Photo | 39 | May 22, 2026 |
|---|
| Jailbreak Attack | JailbreakBench | 39 | May 21, 2026 |
|---|
| ICD classification | psychiatric dataset (test) | 39 | May 21, 2026 |
|---|
| ICD classification | psychiatric dataset (val) | 39 | May 21, 2026 |
|---|
| Node Classification | OGBN-ArXiv standard (test) | 39 | May 29, 2026 |
|---|
| Action Recognition | P3AR-NTU | 39 | May 20, 2026 |
|---|
| Natural Language Understanding and Reasoning | Standard Suite Zero-shot (PiQA, ARC-E, ARC-C, HellaSwag, WinoGrande, BoolQ) | 39 | May 19, 2026 |
|---|