Utility Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
4.83Just-Eval Average Score
50
May 12, 2026
49.5MMLU Score
45
May 15, 2026
78.4Utility
32
Apr 14, 2026
68.85Score
24
Mar 4, 2026
93.6Pass@1 Accuracy
22
May 7, 2026
53Accuracy (pass@1)
22
May 7, 2026
78Accuracy (pass@1)
22
May 7, 2026
5.13Delta NQ-Open
17
Mar 20, 2026
0.2ΔMMLU
17
Mar 20, 2026
70.4MMLU Accuracy
16
May 26, 2026
5.24Anchor-PPL
16
Mar 20, 2026
8.08Utility Score
15
Jun 16, 2026
66.6Balanced Acc
13
Feb 26, 2026
68.6Balanced Acc
13
Feb 26, 2026
67.3Balanced Acc
13
Feb 26, 2026
60.3Balanced Acc
13
Feb 26, 2026
81.8Balanced Accuracy
13
Feb 26, 2026
73.2Accuracy
13
Feb 26, 2026
92.3Score
13
Feb 26, 2026
63.4Average Accuracy
12
Feb 26, 2026
72.1Balanced Acc
11
Feb 26, 2026
9.12Utility Score
8
Apr 20, 2026
87.02MMbench Score
7
Feb 26, 2026
95.9Utility
4
Jun 24, 2026
97.2Compliance
3
Apr 10, 2026