Multimodal Misinformation Detection on MMFakeBench (val)
81.65F1 ScoreMERIT
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MERITModel=GPT-4o-mini2025.10 | 81.65 | 75.1 | 84.3 | 79.14 | 65.67 | |
| MMD-AgentModel=GPT-4o-mini2025.10 | 78.8 | 72.5 | 85.59 | 73 | 71.33 | |
| GPT-4V MMD-AgentModel=GPT-4V2025.10 | 74 | 76.8 | 73.4 | 75.5 | — | |
| GPT-4V StandardModel=GPT-4V2025.10 | 72.3 | 75.6 | 72.1 | 72.8 | — |