Natural Language Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
89.62Accuracy
43
May 25, 2026
82.5Average Score
29
Feb 26, 2026
54Synth. Reason. (AS)
16
Feb 26, 2026
75.2BoolQ Accuracy
4
Apr 14, 2026
54.4Accuracy
4
Feb 26, 2026