STEM Question Answering
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
76.2Correctness
25
Jun 11, 2026
45First-Token Accuracy
24
Apr 6, 2026
98.3First-Token Accuracy
24
Apr 6, 2026
93.1Score
14
Feb 26, 2026
0.181Accuracy
5
Feb 26, 2026