Multi-image Spatial Reasoning on MindCube-Tiny (test)
69.2Aro AccuracyGPT-5.2
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-5.2Category=Proprietary2026.02 | 69.2 | 95.5 | 41.5 | 58.4 | |
| HATCHBase Model=Qwen2.5-VL-3B2026.02 | 65.6 | 31 | 50.2 | 50.2 | |
| CAMCUEBackbone=Qwen2.5-VL-7B, Evaluation Mode=Pose-Only (without imagined image feedback)2026.02 | 63.2 | 75 | 31.67 | 47.43 | |
| GPT-4.1Category=Proprietary2026.02 | 59.6 | 49.5 | 47.2 | 50.6 | |
| InternVL-2.5-4BCategory=Open-Weight2026.02 | 57.6 | 33.5 | 45.2 | 45.9 | |
| SpatialLadder-3BBase Model=Qwen2.5-VL-3B2026.02 | 56.4 | 31.5 | 47.8 | 46.8 | |
| Spatial-MLLM-4BBase Model=Qwen2.5-VL-3B2026.02 | 52 | 29 | 34.5 | 37.6 | |
| Gemini-3-ProCategory=Proprietary2026.02 | 51.6 | 54.5 | 46 | 49 | |
| LLaVA-OneVision-4BCategory=Open-Weight2026.02 | 50 | 33.5 | 39.2 | 40.7 | |
| Qwen2.5-VL-3BBase Model=Qwen2.5-VL-3B2026.02 | 46.4 | 34.5 | 35.3 | 37.8 | |
| Qwen2.5-VL-72BCategory=Open-Weight2026.02 | 44.8 | 42 | 41.2 | 42.2 | |
| Video-R1Base Model=Qwen2.5-VL-7B2026.02 | 44 | 30 | 38.5 | 38.2 | |
| Qwen2.5-VL-32BCategory=Open-Weight2026.02 | 43.6 | 41.5 | 38.2 | 40 | |
| Qwen2.5-VL-7BBase Model=Qwen2.5-VL-7B2026.02 | 38 | 33.5 | 32.7 | 34.1 | |
| SpaceR-7BBase Model=Qwen2.5-VL-7B2026.02 | 31.6 | 31.5 | 28.7 | 29.9 | |
| InternVL-2.5-8BCategory=Open-Weight2026.02 | 25.6 | 37 | 20.3 | 24.8 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B, Evaluation Mode=Base (backbone-only)2026.02 | 21.4 | 38.7 | 29.5 | 29.3 |