Our new X account is live! Follow @wizwand_team for updates

Spatial Reasoning on Visual Spatial Reasoning (VSR)

95.4Accuracy

Human Estimate

Updated 4d ago

Evaluation Results

Method	Links
Human Estimate 2025.04		95.4
Qwen2.5-VL 2026.01		89.116
SpaRE-7B 2025.04		85.4
GPT-5-mini 2025.12		84.4
Qwen3-VL-4B-Thinking + DiG 2025.12		83.9
Qwen3-VL-8B-Thinking + DiG 2025.12		83
Qwen3-VL-8B-Thinking 2025.12		82.6
Qwen3-VL-4B-Thinking 2025.12		82.5
GPT-5-mini 2025.12		82.4
Qwen2VL-7B 2025.04		82.3
SpaRE-2B 2025.04		80.8
GPT-4o 2025.04		79
LLaVA-OneVision 2026.01		77.741
Qwen2.5-VL-7B 2025.12		77.7
Molmo 2026.01		76.432
VALOR 2025.12		75.6
GPT-4o 2025.02		74.8
GPT-4o-mini 2025.04		74
InternVL2-8B 2025.04		73.1
LLaVA-NeXT-8B 2025.04		71.9
SmolVLM2 2026.01		71.277
Qwen2VL-2B 2025.04		70.3
InternVL2-2B 2025.04		68.7
SpaceLLaVa 2025.04		65.9
PaliGemma 2026.01		65.139
Magma-8B (Full) 2025.02		65.1
LLaVA-NeXT 2026.01		63.994
Magma-8B (Act) 2025.02		62.8
LLaVA-AIMv2-2D-RoPE 2026.01		60.311
Magma-8B (Full) 2025.02		58.1
LLaVA-SigLIP-2D-RoPE 2026.01		57.692
LLaVA-2D-RoPE 2026.01		57.201
LLaVA-1.5-7B 2025.02		57.1
LLaVA-SigLIP2-2D-RoPE 2026.01		56.465
LLaVA-AIMv2 2026.01		56.219
LLaVA v1.5 2026.01		55.81
Gemma3-4b-it 2026.01		55.074
LLaVA-SigLIP 2026.01		54.91
CLIP-ViT-H-14 2026.01		54.5
CLIP-ViT-B/32 2026.01		53.7
MiniCPM-V 4.5 2026.01		53
LLaVA-SigLIP2 2026.01		52.701
LLaVA-1.6-7B 2025.02		52.2
InternVL3.5 8B 2026.01		52
InternVL3.5 8B 2026.01		51.6
MiniCPM-V 4.5 2026.01		51.4
Random 2025.04		50
Random 2026.01		50