List your company

Multi-image Spatial Reasoning on SPAR-Bench-MV (test)

43.7Score (Low Difficulty)GPT-5.2
12.18820.36928.5536.731Feb 9, 2026

Evaluation results

MethodLinks
GPT-5.2Category=Proprietary2026.0243.741.566.452.6
HATCHBase Model=Qwen2.5-VL-3B2026.0241.347.467.153.6
Gemini-3-ProCategory=Proprietary2026.0240.621.75843.1
GPT-4.1Category=Proprietary2026.0238.23743.139.9
InternVL-2.5-8BCategory=Open-Weight2026.0231.830.848.238.3
SpaceR-7BBase Model=Qwen2.5-VL-7B2026.0231.632.349.239.1
InternVL-2.5-4BCategory=Open-Weight2026.0230.330.138.533.6
Video-R1Base Model=Qwen2.5-VL-7B2026.0229.43046.236.5
Qwen2.5-VL-72BCategory=Open-Weight2026.0226.932.743.635.4
SpatialLadder-3BBase Model=Qwen2.5-VL-3B2026.0226.233.544.535.8
Qwen2.5-VL-32BCategory=Open-Weight2026.0225.62945.434.7
LLaVA-OneVision-4BCategory=Open-Weight2026.02232840.131.5
Spatial-MLLM-4BBase Model=Qwen2.5-VL-3B2026.0222.332.435.230.4
Qwen2.5-VL-7BBase Model=Qwen2.5-VL-7B2026.0216.929.440.730.1
Qwen2.5-VL-3BBase Model=Qwen2.5-VL-3B2026.0213.426.232.624.9