3D Spatial Reasoning on Omni3D-Bench (test)
65.3Yes/No AccGPT-4o
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-4oCategory=Generic VLMs2025.12 | 65.3 | 60.5 | 18.6 | 26.7 | 8.2 | 27.2 | |
| LLaVA-OV-7B-ChatCategory=Generic VLMs2025.12 | 60 | 27.9 | 22.9 | 26.8 | 11.1 | 23 | |
| SpatialBot-3BCategory=Spatial-Finetuned VLMs2025.12 | 60 | 30.2 | 0 | 17.7 | 8.5 | 18.8 | |
| TVPCategory=Visual Programming, Tool Library Usage=Enabled2025.12 | 60 | 61.6 | 24.3 | 36.5 | 19.3 | 33.3 | |
| TVP w/o Tool LibCategory=Visual Programming, Tool Library Usage=Disabled2025.12 | 60 | 61.6 | 21.4 | 35.5 | 17 | 31.7 | |
| Qwen2-VL-7B-InstCategory=Generic VLMs2025.12 | 58.7 | 33.7 | 12.9 | 21.5 | 10 | 21.8 | |
| ViperGPTCategory=Visual Programming, Source=VADAR paper2025.12 | 56 | 42.4 | 20 | 15.4 | — | — | |
| VADARCategory=Visual Programming, Source=VADAR paper2025.12 | 56 | 57.6 | 21.7 | 35.5 | 15.9 | 29.9 | |
| VisProgCategory=Visual Programming, Source=VADAR paper2025.12 | 54.7 | 25.9 | 2.9 | 0.9 | — | — | |
| SpaceMantisCategory=Spatial-Finetuned VLMs2025.12 | 53.3 | 30.2 | 4.3 | 21.4 | 8.2 | 18.2 | |
| Molmo-7B-DCategory=Generic VLMs2025.12 | 46.7 | 41.9 | 18.6 | 28.4 | 8.9 | 21.6 |