Multi-image in the Wild on Mantis
77.6AccuracyLLaVA-OneVision-72B
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVA-OneVision-72BModel Scale=72B2024.08 | 77.6 | |
| LLaVA-OneVision-7BModel Scale=7B2024.08 | 64.2 | |
| GPT-4VVersion=V-Preview2024.08 | 62.7 | |
| LLaVA-OneVision-0.5BModel Scale=0.5B2024.08 | 39.6 |