Visual Reasoning on Sem. Corr.
58.3AccuracyGPT-4o
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4oVisual Sketchpad=Enabled2024.06 | 58.3 | |
| GPT-4oVisual Sketchpad=Disabled2024.06 | 48.6 | |
| GPT-4 TurboVisual Sketchpad=Enabled2024.06 | 42.4 | |
| SOTA [49]2024.06 | 33.1 | |
| LLaVA-1.5-13B2024.06 | 32.4 | |
| GPT-4 TurboVisual Sketchpad=Disabled2024.06 | 30.9 | |
| GPT-4V-preview2024.06 | 28.8 | |
| Gemini-Pro2024.06 | 26.6 | |
| Claude 3 OPUS2024.06 | 25.2 | |
| LLaVA-NeXT-34B2024.06 | 23.7 | |
| LLaVA-1.5-7B2024.06 | 23 |