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 |
|---|---|---|---|---|---|
| Facial Forgery Detection | AM IAFaces (test) | 30 | May 12, 2026 | ||
| Molecular Property Optimization | QM9 frontier-orbital (test) | 30 | May 12, 2026 | ||
| Multilingual Extractive Question Answering | Multilingual QA matrix (JSQuAD, TyDi-QA, and XQuAD) binary feedback at Q ≈ 10^6 (test) | 30 | May 12, 2026 | ||
| Retrosynthesis planning | Pistachio Hard | 30 | May 12, 2026 | ||
| Retrosynthesis planning | Pistachio Reachable |
| 30 |
| May 12, 2026 |
| Retrosynthesis planning | USPTO Easy | 30 | May 12, 2026 |
|---|
| Time Series Forecasting Uncertainty Quantification | GIFT-Eval Bench22 | 30 | May 12, 2026 |
|---|
| Time Series Forecasting Uncertainty Quantification | GIFT-Eval Bench10 | 30 | May 12, 2026 |
|---|
| Multi-Tote Handling | 2D Multi-Tote Handling Robotic Systems (Hairobotic) Extended Large-Scale Instances | 30 | May 12, 2026 |
|---|
| Clustering | Shuttle | 30 | Jun 18, 2026 |
|---|
| General instruction following | IFBench | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Goodreads | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Sports | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Pet | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Baby | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Office | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Auto | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Elec | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Goodread | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Sports | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Music | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Pet | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Baby | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Office | 30 | May 12, 2026 |
|---|
| Relation-Conditioned Retrieval | Auto | 30 | May 12, 2026 |
|---|