Multi-image Spatial Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
51Overall Score
32
Jul 2, 2026
80.4Accuracy
19
May 26, 2026
69.2Aro Accuracy
17
Feb 26, 2026
38Overall Accuracy
15
Feb 26, 2026
43.7Score (Low Difficulty)
15
Feb 26, 2026