Spatial Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
75.33Accuracy
7
Mar 17, 2026
93.48Accuracy
7
Mar 17, 2026
80.21Accuracy
7
Mar 17, 2026
73Accuracy
7
Feb 26, 2026
44Accuracy
7
Feb 26, 2026
61.4Accuracy
7
Feb 26, 2026
98.4Accuracy
6
Jun 30, 2026
0.984Accuracy
6
Jun 30, 2026
72.1NQ Score
6
Jun 18, 2026
71.22D Translation
6
Jun 9, 2026
67.5Accuracy
6
May 18, 2026
79.3Relative Position Accuracy
6
Apr 24, 2026
73.3HR-8K Average Accuracy
6
Apr 23, 2026
84.5Accuracy
6
Mar 30, 2026
95Accuracy (What'sUp Test)
6
Mar 24, 2026
41.85LLM-Judge Accuracy
6
Mar 24, 2026
81.89Spatial Score
6
Mar 24, 2026
53.64Accuracy
6
Feb 26, 2026
63.67Accuracy (zero-shot)
6
Feb 26, 2026
44.1NL-to-Format Accuracy
6
Feb 26, 2026
44.3Position IoU
6
Feb 26, 2026
74.1Accuracy (All)
6
Feb 26, 2026
76.5Accuracy (All)
6
Feb 26, 2026
99.88Task 17 Accuracy
6
Feb 26, 2026
55.1Accuracy
5
Apr 23, 2026