| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Image Captioning | CapArena-Auto | Average Win-Rate Margin91.7 | 29 | |
| Open-ended Generation | CapArena | Average Score11.83 | 7 | |
| Image Captioning | CapArena | Claude Score41.67 | 6 | |
| Image Captioning | CapArena (test) | GPT-Score19.15 | 5 |