Text-to-Image Generation Evaluation on RichHF-18K
40.4Pearson-rUnifiedReward_Q
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UnifiedReward_QScale=8B2025.06 | 40.4 | — | |
| UnifiedReward_LScale=7B2025.06 | 39.9 | — | |
| Gemini-2.5-ProScale=/2025.06 | 39.7 | — | |
| Qwen3-VLScale=8B2025.06 | 38.9 | — | |
| MinosScale=8B2025.06 | 36 | — | |
| LLaVA-CriticScale=72B2025.06 | 33 | — | |
| GPT-4oScale=/2025.06 | 31.1 | — | |
| LLaVA-OVScale=72B2025.06 | 27.2 | — | |
| PickScore2025.06 | 22.4 | 18.3 | |
| HPS v22025.06 | 19.5 | 13.1 | |
| LLaVA-CriticScale=7B2025.06 | 18.4 | — | |
| Prometheus-VScale=7B2025.06 | 8.19 | — | |
| LLaVA-OVScale=7B2025.06 | 5.85 | — |