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 |
|---|---|---|---|---|---|
| Error Attribution | GSM8k Left Dense | 30 | May 11, 2026 | ||
| Long-context understanding | LongBench | 30 | May 11, 2026 | ||
| Minimize transmit power with CI constraints and imperfect CSI | Hybrid Precoding Robustness Scenario 09 | 30 | May 8, 2026 | ||
| Maximize sum rate with interference temperature constraints | Cognitive Radio Inference Robustness Scenario 07 | 30 | May 8, 2026 | ||
| Maximize CI margin with 1-bit DAC constraints | Hybrid 1-bit MISO Scenario 03 |
| 30 |
| May 8, 2026 |
| Maximize CI margin with unit modulus constraints | Constant Envelope precoding Scenario 02 | 30 | May 8, 2026 |
|---|
| Minimize transmit power with rate constraints | MU-MIMO Scenario 01 | 30 | May 8, 2026 |
|---|
| Mathematical reasoning | AIME24 | 30 | May 22, 2026 |
|---|
| Rainfall Estimation | Gridded Rainfall Products (test) | 30 | May 8, 2026 |
|---|
| Question Answering | MultiFieldQA-zh | 30 | May 8, 2026 |
|---|
| Retrieval | QASPER (test) | 30 | May 8, 2026 |
|---|
| Square Nut Assembly | Robomimic Square-State | 30 | May 7, 2026 |
|---|
| Can Pick-and-Place | Robomimic Can-State | 30 | May 7, 2026 |
|---|
| Lift | Robomimic Lift-State | 30 | May 7, 2026 |
|---|
| Mathematical Reasoning | Math Reasoning AIME24, AIME25, HMMT25 | 30 | Jun 23, 2026 |
|---|
| RGB-T Pedestrian Detection Attack | FLIR aligned (test) | 30 | May 7, 2026 |
|---|
| Irregular Time-series Modeling | Spiral | 30 | May 7, 2026 |
|---|
| Trait Alignment | Persona Vectors | 30 | May 6, 2026 |
|---|
| ECG Denoising | Clean canine ECG signals | 30 | May 6, 2026 |
|---|
| Object Manipulation | OGBench cube play (Single) | 30 | Jun 9, 2026 |
|---|
| Law Retrieval | Austrian-law 5,000-query (test) | 30 | May 6, 2026 |
|---|
| Path Planning | Mixed Simulation Terrain (test) | 30 | May 4, 2026 |
|---|
| Multi-Depot Vehicle Routing Problem | MDVRP n=100 | 30 | May 1, 2026 |
|---|
| Joint image compression and wireless transmission | CelebA TDL-A channels | 30 | May 1, 2026 |
|---|
| Sensor Selection | Weather Forecast 180 spatial ablation configurations (2 models × 5 cities × 3 variables × 3 patches × 2 perturbations) | 30 | May 1, 2026 |
|---|