Natural Language Inference
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
94.7Accuracy
694
May 14, 2026
93.5Accuracy
590
May 28, 2026
100Accuracy
218
Jun 4, 2026
90Average Accuracy
167
Jun 2, 2026
99.7Accuracy
154
Feb 26, 2026
87.1Accuracy
131
Jun 18, 2026
98.2Accuracy
129
Jun 1, 2026
91.7Accuracy
110
Feb 26, 2026
74.02Accuracy
107
Jun 2, 2026
96.56Accuracy
93
Jun 4, 2026
86.57Accuracy
89
Feb 26, 2026
96.8Accuracy
86
May 14, 2026
91.58Accuracy
85
May 29, 2026
88.05Average Worst-Group Accuracy
81
Apr 21, 2026
90.8Accuracy (matched)
80
Jun 8, 2026
93.6Accuracy
71
Feb 27, 2026
91Accuracy
68
Feb 26, 2026
85.38Accuracy
65
Feb 26, 2026
67.9Accuracy
64
Feb 26, 2026
66.5Accuracy
64
Feb 26, 2026
79.1Accuracy
60
Feb 26, 2026
77Accuracy
57
Feb 26, 2026
72.46Accuracy (Chinese)
56
Feb 26, 2026
81.4Accuracy
56
Feb 26, 2026
78.65Accuracy
54
Feb 26, 2026