List your company

Spatial Logical Reasoning on SpatiaLQA

97.6Rchuman
21.05640.92860.880.672Feb 24, 2026

Evaluation results

MethodLinks
human2026.0297.697.697.692.392.792.5
GPT-52026.0267.886.47639.258.547
Gemini-2.5-pro2026.0265.386.274.33153.639.3
GPT-4.12026.0265.284.273.530.251.338
GPT-5-mini2026.0264.586.673.932.856.241.2
Qwen-VL-Max2026.02628370.925.645.232.7
Gemini-2.5-flash2026.026288.672.929.557.338.9
Qwen2.5-VL-72B-InsSize=72B, instruction-tuned=true2026.026082.969.623.944.831.2
Claude-4-sonnet2026.0259.48268.927.852.136.3
GPT-4o2026.025978.567.4193725.1
GPT-4o-mini2026.0254.376.463.514.126.418.4
GPT-4.1-mini2026.0253.887.166.521.251.430
MiniCPM-V-4.5Size=9B2026.025059.454.210.918.613.7
InternVL3.5-14B-InsSize=14B, instruction-tuned=true2026.0249.366.856.81424.917.9
Cosmos-Reason1Size=7B2026.0248.266.956.111.940.318.3
PixtralSize=12B2026.0248.1705713.731.219
Claude-3-7-sonnet2026.0247.180.259.31952.327.9
LLaVA1.6-vicuna-7BSize=7B2026.024748.647.81015.512.2
GLM-4.1V-9B-BaseSize=9B2026.0245.678.157.612.131.217.5
Gemma3-27B-InsSize=27B, instruction-tuned=true2026.0244.160.5519.920.613.4
Qwen3-VL-8B-InsSize=8B, instruction-tuned=true2026.024465.752.713.436.219
GLM-4.1V-9B-ThinkingSize=9B, thinking-mode=true2026.024482.757.512.437.118.6
DeepSeek-VL2-SmallSize=16B2026.0243.484.957.410.559.117.9
InternVL3.5-8B-InsSize=8B, instruction-tuned=true2026.0243.171.553.813.433.619.2
LLaVA1.6-vicuna-13BSize=13B2026.024248.54511.924.916.1
Qwen2.5-VL-32B-InsSize=32B, instruction-tuned=true2026.024275.5549.224.613.4
InternVL3.5-4B-InsSize=4B, instruction-tuned=true2026.024267.851.911.330.516.5
LLaVA1.6-mistralSize=7B2026.0241.547.344.27.213.69.4
BLIP2-Flan-T5-xxlSize=12B2026.0241.367.751.300.20
Gemma3-12B-InsSize=12B, instruction-tuned=true2026.024164.650.29.124.913.4
LLaVA1.5-7BSize=7B2026.0240.339.8406.497.5
LLaVA1.5-13BSize=13B2026.0239.149.442.59.119.312.3
Qwen3-VL-4B-InsSize=4B, instruction-tuned=true2026.0238.878.85212.546.519.6
Qwen-VL-Plus2026.0238.386539.537.515.1
Phi3.5-vision-InsSize=4B, instruction-tuned=true2026.0236.638.437.55.510.37.2
SpaceOmSize=4B2026.0234.171.246.18.735.113.9
Qwen2.5-VL-3B-InsSize=3B, instruction-tuned=true2026.0232.575.845.53.37.44.6
Qwen2.5-VL-7B-InsSize=7B, instruction-tuned=true2026.0232.573.245.13.315.55.5
Kimi-VL-A3B-ThinkingSize=16B, thinking-mode=true2026.023292.247.54.442.38
Kimi-VL-A3B-InsSize=16B, instruction-tuned=true2026.0228.778.2421.111.42.1
BLIP2-OPTSize=4B2026.0224.699.939.500—
BLIP2-Flan-T5-xlSize=4B2026.022474.336.30.10.20.1