WISE
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| World Knowledge Image Generation | WISE | Overall Score89 | 93 | |
| Knowledge-grounded reasoning | WISE | Overall Score89 | 68 | |
| Multimodal Understanding and Generation | WISE | Overall Accuracy80 | 62 | |
| Text-to-Image Generation | WISE | Cultural Score81 | 48 | |
| Text-to-Image Generation | WISE | WISE Score0.7 | 48 | |
| Reasoning-based text-to-image generation | WISE | Overall Score80 | 33 | |
| Text-to-Image Generation | WISE (test) | Overall Score80 | 32 | |
| Reasoning Generation | WISE 1.0 (test) | Overall Score80 | 17 | |
| World Knowledge Reasoning | WISE random subset of 200 samples | Cultural Accuracy94 | 17 | |
| Text-to-image generation | WISE | Culture Score0.76 | 15 | |
| Text-to-image generation | WISE 48 (test) | Cultural Score0.81 | 12 | |
| Knowledge-informed text-to-image generation | WISE 96 | Cultural Score63 | 10 | |
| Text-to-Image Generation | WISE | Physics Score67 | 9 | |
| World Knowledge-aware Image Generation | WISE | Consistency1.26 | 2 |
Showing 14 of 14 rows