Calibration
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
0.01Top-1 ECE
4
Mar 27, 2026
7.5ECE
4
Mar 23, 2026
0.794ECE (Wrong Samples)
4
Feb 26, 2026
0.586ECE (Wrong Samples)
4
Feb 26, 2026
0.236ECE
4
Feb 26, 2026
0.277ECE
4
Feb 26, 2026
1.1Friedman Test Statistic
4
Feb 26, 2026
0.04ECE (Noise-level 1)
4
Feb 26, 2026
0.087ECE
3
May 4, 2026
0.187ECE
2
Feb 26, 2026
0.179ECE
2
Feb 26, 2026
0.226ECE
2
Feb 26, 2026
0.104ECE
2
Feb 26, 2026
0.113ECE
2
Feb 26, 2026
0.116ECE
2
Feb 26, 2026
—
—Primary metric
0
Feb 26, 2026