Spatial & 3D Understanding on Ego3DBench (OOD)
59.93Accuracy (ACC)Qwen3-VL-32B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-VL-32BModel Category=Open-Source Models2026.04 | 59.93 | 15.7 | |
| Qwen2.5-VL-32BModel Category=Open-Source Models2026.04 | 57.3 | 15.87 | |
| XEmbodied (Best)Model Category=Our Model, Zero-shot evaluation=true2026.04 | 55.28 | 9.25 | |
| Mimo-EmbodiedModel Category=Embodied Models2026.04 | 53.62 | 16.57 | |
| Gemini-1.5Model Category=Proprietary Models2026.04 | 53.5 | 19.62 | |
| Qwen3-VL-A3B-30BModel Category=Open-Source Models2026.04 | 53.15 | 13.84 | |
| GPT-4oModel Category=Proprietary Models2026.04 | 52.7 | 19.2 | |
| DriveMMModel Category=Embodied Models2026.04 | 49.73 | 12.68 | |
| UniVG-R1-7BModel Category=Spatial Models2026.04 | 46.82 | 53.85 | |
| Cosmos-R1Model Category=Embodied Models2026.04 | 45.62 | 23.41 | |
| Qwen2.5-VL-7BModel Category=Open-Source Models2026.04 | 41.1 | 30.36 | |
| PR1-Counting-2BModel Category=Spatial Models2026.04 | 40.32 | 51.95 | |
| PR1-Grounding-2BModel Category=Spatial Models2026.04 | 39.54 | 12.34 | |
| PR1-OCR-2BModel Category=Spatial Models2026.04 | 38.88 | 11.6 | |
| PR1-Detection-3BModel Category=Spatial Models2026.04 | 36.74 | 62.61 |