Out-of-view Understanding on OpenView-Bench Directional
0.85Choice AccuracyHuman-level
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Human-level2025.12 | 0.85 | — | — | |
| MiniCPM-V 4.5Rank=32025.12 | 0.5861 | 0.8479 | 0.497 | |
| Gemini-2.5-flashRank=22025.12 | 0.574 | 0.9263 | 0.5317 | |
| GPT-5Rank=12025.12 | 0.5574 | 0.9892 | 0.5514 | |
| Qwen3-VL-8B-ThinkingRank=5, Thinking Mode=true2025.12 | 0.5514 | 0.7616 | 0.4199 | |
| GPT-4oRank=92025.12 | 0.5498 | 0.5879 | 0.3233 | |
| Qwen2.5-VL-72B-InstructRank=62025.12 | 0.5483 | 0.8072 | 0.4426 | |
| Qwen2.5-VL-32B-InstructRank=42025.12 | 0.5468 | 0.8398 | 0.4592 | |
| GPT-4o-miniRank=102025.12 | 0.5106 | 0.6509 | 0.3323 | |
| InternVL3.5-8BRank=13, Thinking Mode=true2025.12 | 0.4909 | 0.48 | 0.2356 | |
| Qwen2.5-VL-3B-InstructRank=172025.12 | 0.4804 | 0.0252 | 0.0121 | |
| Qwen3-VL-8B-InstructRank=82025.12 | 0.4728 | 0.9073 | 0.429 | |
| InternVL3.5-8BRank=142025.12 | 0.4713 | 0.5 | 0.2356 | |
| GLM-4.1V-9B-ThinkingRank=11, Thinking Mode=true2025.12 | 0.4698 | 0.6881 | 0.3233 | |
| Qwen2.5-VL-7B-InstructRank=152025.12 | 0.4683 | 0.5097 | 0.2387 | |
| GLM-4.1V-9B-BaseRank=132025.12 | 0.4653 | 0.5065 | 0.2356 | |
| Ovis2.5-9BRank=72025.12 | 0.4592 | 0.8125 | 0.3731 | |
| LLaVA-NeXT-Mistral-7BRank=162025.12 | 0.3142 | 0.2837 | 0.0891 |