Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Multitask Language Understanding on MMLU (Accuracy)

89.8Accuracy

Average human expert performance

53.03662.580572.12581.6695Dec 8, 2021Aug 17, 2022Apr 26, 2023Jan 4, 2024Sep 12, 2024May 22, 2025Jan 30, 2026
Updated 2d ago

Evaluation Results

MethodLinks
89.8
2025.07
77.82
2025.07
77.82
2026.01
77
2026.01
76.9
2026.01
76.5
2026.01
76.2
2025.07
74.6
2025.12
74.27
2025.12
74.19
2025.12
74.16
2025.07
74.06
2025.07
74.06
2025.07
74.04
2025.12
73.81
2025.12
73.61
2025.12
73.59
2026.01
73.5
2026.01
73.5
2025.12
73.05
2025.12
73.02
2025.12
72.74
2025.12
72.72
2025.12
72.64
2025.12
72.55
2025.12
72.33
2025.12
72.31
2025.07
72.25
2025.12
72.22
2025.12
72.06
2025.12
71.97
2025.07
71.64
2026.01
71.6
2026.01
71.4
2025.07
70.99
2026.01
70.1
2026.01
69.9
2025.07
69.46
2025.07
69.43
2025.12
68.81
2025.07
67.3
2025.07
67.3
2026.01
67.1
2026.01
67
2025.12
66.7
2025.12
66.3
2025.12
66.2
2025.12
66.1
2025.12
65.7
2026.01
65.4
2026.01
65.4
2025.12
65.2
2025.12
65.13
2025.07
65.02
2025.12
65
63.4
2024.09
63.2
2025.12
62.4
2024.09
62.4
2026.01
62.23
2025.07
62
2025.07
61.37
2026.01
61.06
2025.12
60.83
2026.01
60.78
2026.01
60.78
2026.01
60.4
2026.01
60.3
2026.01
60.04
2026.01
60.04
2025.12
60.02
2025.12
60
2021.12
60
2025.12
59.96
2025.12
59.84
2025.12
59.84
2025.07
59.75
2025.07
59.65
2025.12
59.54
2025.12
59.49
2025.12
59.31
2025.12
59.19
2025.12
59.09
2026.01
58.95
2026.01
58.11
2026.01
58.11
2025.12
57.66
57.1
2025.07
56.3
2025.07
55.84
2025.12
55.25
2025.07
55.23
2026.01
55
2025.12
54.97
2025.07
54.96
2025.07
54.96
2026.01
54.9
2025.12
54.77
2025.07
54.58
2025.12
54.45
Showing 100 of 206 rows