Long T2I Alignment Interpreting on LongT2IBench (test)
80.2Entity Align AccLongT2IExpert
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LongT2IExpert2025.12 | 80.2 | 36.5 | 71.9 | 53.8 | 18.9 | 47.3 | 40.6 | 20.8 | 35.2 | 60.7 | 25.8 | 53.2 | |
| Qwen2.5-v1-max2025.12 | 49.9 | 22.8 | 44.8 | 28.2 | 10.8 | 24.9 | 9.8 | 4.7 | 8.4 | 31.7 | 12.6 | 27.7 | |
| Gemini-1.5-pro2025.12 | 49.2 | 36.1 | 46.7 | 31.6 | 16.6 | 28.8 | 13.6 | 12.8 | 13.4 | 33.6 | 21.9 | 31.1 | |
| Gemini-1.5-flash2025.12 | 48.4 | 23.2 | 43.6 | 26.4 | 9.2 | 23.2 | 6.4 | 3.2 | 5.6 | 29.7 | 11.8 | 25.9 | |
| GPT-4o2025.12 | 44.2 | 30.7 | 41.6 | 27.6 | 13.9 | 25 | 10 | 11 | 10.2 | 29.3 | 18.6 | 27 | |
| GPT-4-Turbo2025.12 | 43.1 | 27.1 | 40.1 | 24.9 | 11.9 | 22.5 | 5.6 | 5 | 5.4 | 26.8 | 14.6 | 24.2 | |
| Grok-32025.12 | 42.9 | 44 | 43.1 | 23 | 16.4 | 21.7 | 6.2 | 10.7 | 7.4 | 26.3 | 23.8 | 25.7 | |
| LLava-v1.62025.12 | 33.6 | 6.5 | 28.4 | 11.4 | 1.9 | 9.6 | 1.9 | 0.3 | 1.5 | 17.6 | 2.9 | 14.5 |