ResearchBenchmarksSpatial Reasoning on Spatial-GridFollow84.5AccuracyGPT-4o23.1439.075570.93Nov 14, 2025Evaluation ResultsMethodMethodLinksAccuracyGPT-4oInference Scheme=Binar...Inference Scheme=Binary verification2025.1184.5GPT-4oInference Scheme=MCQ (...Inference Scheme=MCQ (5× MV)2025.1178.7GPT-4oInference Scheme=MCQInference Scheme=MCQ2025.1178.3CogVLMInference Scheme=Binar...Inference Scheme=Binary verification2025.1131.3CogVLMInference Scheme=MCQ (...Inference Scheme=MCQ (5× MV)2025.1125.7CogVLMInference Scheme=MCQInference Scheme=MCQ2025.1125.5