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 |
|---|---|---|---|---|---|
| Object Detection | COCO | 24 | Jul 8, 2026 | ||
| Referring Object Detection | RefCOCO low-data (testB) | 24 | Mar 26, 2026 | ||
| Referring Object Detection | RefCOCO low-data (testA) | 24 | Mar 26, 2026 | ||
| Referring Object Detection | RefCOCO low-data (val) | 24 | Mar 26, 2026 | ||
| Referring Image Segmentation | RefCOCO RefCOCO+ RefCOCOg Overall Average |
| 24 |
| Mar 20, 2026 |
| Object Hallucination Mitigation | COCO 2014 | 24 | Mar 18, 2026 |
|---|
| Conditional generation | MS-COCO 512x512 (val) | 24 | Mar 13, 2026 |
|---|
| Conditional Image Generation | MS-COCO 512 x 512 1.5 | 24 | Mar 13, 2026 |
|---|
| Text-to-Image Generation | COCO 512x512 | 24 | Mar 13, 2026 |
|---|
| Object Hallucination Evaluation | MSCOCO POPE (random popular adversarial) | 24 | Mar 12, 2026 |
|---|
| Object Detection | COCO mini 10% (val) | 24 | Mar 11, 2026 |
|---|
| Image-Text Retrieval | MS-COCO 1K (test) | 24 | Mar 9, 2026 |
|---|
| Jailbreaking | COCO-I2VSafetyBench | 24 | Feb 26, 2026 |
|---|
| Image Captioning | MS-COCO 1,000 images subset 2014 2017 (val) | 24 | Feb 26, 2026 |
|---|
| Image Reconstruction | COCO (test) | 24 | Feb 26, 2026 |
|---|
| Object Hallucination | COCO 512-token budget (test) | 24 | Feb 26, 2026 |
|---|
| Object Hallucination | COCO 64-token budget (test) | 24 | Feb 26, 2026 |
|---|
| Image Retrieval | MSCOCO @5000 (seen classes) | 24 | Feb 26, 2026 |
|---|
| Image Retrieval | MSCOCO @5000 15%-unseen | 24 | Feb 26, 2026 |
|---|
| Multi-label class-incremental learning | MS-COCO B0-C5 | 24 | Mar 9, 2026 |
|---|
| Multi-label class-incremental learning | MS-COCO B20-C4 | 24 | Mar 9, 2026 |
|---|
| Image Generation | MS-COCO 10k (test) | 24 | Mar 4, 2026 |
|---|
| Image Denoising | COCO 2017 (test) | 24 | Feb 26, 2026 |
|---|
| Post-attack image integrity | COCO | 24 | Feb 26, 2026 |
|---|
| Scene Text Recognition | COCO | 24 | Feb 27, 2026 |
|---|