Calibration
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
3.18Expected Calibration Error (ECE)
22
Jun 5, 2026
3.75Expected Calibration Error (ECE)
22
Jun 5, 2026
0.298NLL
21
May 26, 2026
0.08ECE
21
Apr 14, 2026
0.653NLL
20
Jul 3, 2026
0.341NLL
20
Jul 3, 2026
0.39ECE (Oracle Estimate)
20
May 12, 2026
1.499NLL
20
Feb 26, 2026
0.531NLL
20
Feb 26, 2026
2.32Brier Score
19
Feb 26, 2026
3.41ECE
18
Feb 26, 2026
2.61ECE (%)
18
Feb 26, 2026
0.028ECE
18
Feb 26, 2026
1.87ECE
18
Feb 26, 2026
0.018ECE
16
Apr 3, 2026
1.506ECE
16
Feb 26, 2026
4.74ECE
14
Feb 26, 2026
0.28ECE
12
Feb 26, 2026
5.56Expected Calibration Error (ECE)
11
Jun 1, 2026
6.12ECE
11
Jun 1, 2026
1.64ECE
11
Jun 1, 2026
7.79ECE
11
Jun 1, 2026
8.15ECE
11
Feb 26, 2026
0.102ECE
10
Feb 26, 2026
4.268Expected Calibration Error
10
Feb 26, 2026