4D Spatial-Temporal Reasoning on MLLM4D-Bench
7.7Camera Absolute Distance ErrorInternVL2.5-38B
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| InternVL2.5-38BModel Category=Open-Source Models2026.02 | 7.7 | 16.4 | 9.4 | 14.1 | 24 | 17.4 | 14.8 | |
| InternVideo2.5-8BModel Category=Open-Source Models2026.02 | 9.4 | 48.2 | 9.5 | 9.3 | 22.3 | 18.6 | 19.6 | |
| LLaVA-NeXT-Video-7BModel Category=Open-Source Models2026.02 | 22.6 | 26 | 24.6 | 24.3 | 33.4 | 32.6 | 27.3 | |
| InternVL3.5-38BModel Category=Open-Source Models2026.02 | 24.3 | 22.6 | 24 | 40.9 | 40.7 | 37.1 | 31.6 | |
| Qwen2.5-VL-7BModel Category=Open-Source Models2026.02 | 25 | 58 | 24.6 | 25 | 42.5 | 30.6 | 34.3 | |
| InternVL2.5-8BModel Category=Open-Source Models2026.02 | 27.5 | 43.4 | 5.8 | 7.2 | 21 | 16.1 | 20.2 | |
| VLM-3RModel Category=3D Spatial Reasoning Models, Base Model=LLaVA-Video-7B2026.02 | 29 | 56 | 24.6 | 29.2 | 20.4 | 24.6 | 30.6 | |
| Qwen2.5-VL-32BModel Category=Open-Source Models2026.02 | 32.2 | 52 | 31 | 30 | 3.2 | 9.1 | 26.2 | |
| InternVL3.5-8BModel Category=Open-Source Models2026.02 | 34.5 | 55.9 | 33.7 | 34.3 | 41.2 | 36 | 39.3 | |
| GPT-4oModel Category=Proprietary Models2026.02 | 34.8 | 57.6 | 32.7 | 36.7 | 56.7 | 51 | 44.9 | |
| Gemini2.5-FlashModel Category=Proprietary Models2026.02 | 35.2 | 59.4 | 24.1 | 37.8 | 55.3 | 50.6 | 43.8 | |
| Qwen3-VL-8BModel Category=Open-Source Models2026.02 | 36 | 60.4 | 33.8 | 27.2 | 26.4 | 28.2 | 35.3 | |
| Gemini2.5-ProModel Category=Proprietary Models2026.02 | 37.4 | 57.5 | 31.1 | 38.6 | 60.1 | 55 | 46.6 | |
| Qwen3-VL-32BModel Category=Open-Source Models2026.02 | 40.4 | 60.8 | 35.1 | 35.2 | 39.5 | 37 | 41.3 | |
| VG-LLMModel Category=3D Spatial Reasoning Models, Base Model=Qwen2.5-VL-7B2026.02 | 54.9 | 55.7 | 55.6 | 55.7 | 61.8 | 54.3 | 56.3 | |
| MLLM-4DBase Model=Qwen2.5-VL-7B2026.02 | 73.3 | 68.1 | 75.7 | 73 | 70.9 | 60.4 | 70.2 | |
| MLLM-4DBase Model=Qwen3-VL-8B2026.02 | 73.4 | 71.9 | 76.3 | 74.3 | 69.2 | 70.9 | 72.7 |