Multi-image Spatial Reasoning on SPAR-Bench-MV + MindCube-Tiny + MMSI-Bench (test)
51Overall ScoreGPT-5.2
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.2Category=Proprietary2026.02 | 51 | |
| GPT-5.2Model Type=Proprietary2026.07 | 51 | |
| OmniView-Space-R1†Mode=Tool-Integrated2026.07 | 50.6 | |
| OmniView-Space-R1‡Mode=Tool-Distillation2026.07 | 46.5 | |
| Gemini-3-ProCategory=Proprietary2026.02 | 44.9 | |
| Gemini-3-ProModel Type=Proprietary2026.07 | 44.9 | |
| HATCHBase Model=Qwen2.5-VL-3B2026.02 | 43.6 | |
| GPT-4.1Category=Proprietary2026.02 | 40.7 | |
| GPT-4.1Model Type=Proprietary2026.07 | 40.7 | |
| OmniView-Space-ZeroMode=Tool-Integrated2026.07 | 40.1 | |
| OmniView-Space-SFTMode=Tool-Distillation2026.07 | 37.9 | |
| Qwen2.5-VL-72BCategory=Open-Weight2026.02 | 36.5 | |
| InternVL-2.5-4BCategory=Open-Weight2026.02 | 35.5 | |
| SpatialLadder-3BBase Model=Qwen2.5-VL-3B2026.02 | 35.4 | |
| Qwen3-VL-8B-InstructModel Type=Open-source General2026.07 | 35 | |
| Cambrian-S-7BModel Type=Spatial Intelligence2026.07 | 34.9 | |
| InternVL3-8BModel Type=Open-source General2026.07 | 34.8 | |
| Video-R1Base Model=Qwen2.5-VL-7B2026.02 | 34.2 | |
| ViLaSR-7BModel Type=Spatial Intelligence2026.07 | 34.1 | |
| SpatialLadder-3BModel Type=Spatial Intelligence2026.07 | 33.9 | |
| Qwen2.5-VL-32BCategory=Open-Weight2026.02 | 33.8 | |
| LLaVA-OneVision-4BCategory=Open-Weight2026.02 | 33.7 | |
| Qwen3-VL-4B-InstructModel Type=Open-source General2026.07 | 33.1 | |
| SpaceR-7BModel Type=Spatial Intelligence2026.07 | 33 | |
| MindCube-3BModel Type=Spatial Intelligence2026.07 | 32.1 | |
| SpaceR-7BBase Model=Qwen2.5-VL-7B2026.02 | 31.9 | |
| Spatial-MLLM-4BBase Model=Qwen2.5-VL-3B2026.02 | 30.7 | |
| Qwen2.5-VL-7BBase Model=Qwen2.5-VL-7B2026.02 | 30.4 | |
| InternVL-2.5-8BCategory=Open-Weight2026.02 | 30.3 | |
| Qwen2.5-VL-7B-InstructModel Type=Open-source General2026.07 | 30.1 | |
| Spatial-MLLM-4BModel Type=Spatial Intelligence2026.07 | 30 | |
| Qwen2.5-VL-3BBase Model=Qwen2.5-VL-3B2026.02 | 29.4 |