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 |
|---|---|---|---|---|---|
| Molecular Property Prediction | QM9 out-of-sample (test) | 31 | Feb 26, 2026 | ||
| Hand gesture recognition | NVGesture | 31 | May 6, 2026 | ||
| Co-speech gesture generation | BEAT All Speakers 2 | 31 | Mar 30, 2026 | ||
| Meme Classification | PrideMM | 31 | Apr 2, 2026 | ||
| Quantum Chemical Prediction | PCQM4M v2 (test-dev) | 31 | Feb 26, 2026 | ||
| Causal Reasoning |
|---|
| XCOPA (test) |
| 31 |
| Jun 1, 2026 |
| Graph Regression | ZINC Subset | 31 | Feb 26, 2026 |
|---|
| Fine-grained classification | Aircraft (test) | 31 | Jun 8, 2026 |
|---|
| Anomaly Detection | BUSZS | 31 | Feb 26, 2026 |
|---|
| Commonsense reasoning | WinoGrande 1.0 (test) | 31 | Apr 21, 2026 |
|---|
| Machine Translation | IWSLT | 31 | Feb 26, 2026 |
|---|
| End-to-End Autonomous Driving | Bench2Drive | 31 | Mar 17, 2026 |
|---|
| Object Detection | UAVDB (test) | 31 | Feb 26, 2026 |
|---|
| Object Detection | UAVDB (val) | 31 | Feb 26, 2026 |
|---|
| Assistant Response Alignment (Helpfulness and Harmlessness) | HH-RLHF (test) | 31 | Feb 26, 2026 |
|---|
| Language Understanding | MMLU | 31 | Feb 26, 2026 |
|---|
| Semantic Segmentation | Pascal Blended 662 labels augmented (1/16 train) | 31 | Feb 26, 2026 |
|---|
| Watermark Segmentation | Experimental setting S2 (n=2500, K=5) | 30 | Jul 9, 2026 |
|---|
| Inverse Watermarking | Experimental Setting n=500, K=2 S1 (test) | 30 | Jul 9, 2026 |
|---|
| Watermark Segmentation | Gumbel-max watermarking Setting S1 | 30 | Jul 9, 2026 |
|---|
| Watermark Segmentation | Inverse watermarking experimental setting S3 (test) | 30 | Jul 9, 2026 |
|---|
| Watermark segmentation | Permute-and-Flip watermarking Setting S2 | 30 | Jul 9, 2026 |
|---|
| Watermark Segmentation | Google C4 | 30 | Jul 9, 2026 |
|---|
| Watermark Segmentation | Red-Green watermarking Setting S2 | 30 | Jul 9, 2026 |
|---|
| Watermark segmentation | Permute-and-Flip watermarking n=1000, K=5, short gaps S3 (test) | 30 | Jul 9, 2026 |
|---|