Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Multitask Language Understanding on MMLU Redux

84.13Accuracy

T2T

2.188423.461744.73566.0083Jun 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
84.1367.08
2026.06
83.4355.5
2026.06
82.8756.24
2026.06
80.7537.73
2026.06
80.6143.48
2026.06
79.3933.91
2026.06
78.5721.54
2026.06
78.5212.56
2026.06
78.0415.64
2026.06
77.867.95
2026.06
77.6610.18
2026.06
77.3821.79
2026.06
77.066.56
2026.06
75.997.83
2026.06
75.874.5
2026.06
74.599.42
2026.06
73.817.42
2026.06
72.766.64
2026.06
72.119.85
2026.06
72.119.85
2026.06
71.831.16
2026.06
71.831.16
2026.06
70.193.43
2026.06
69.735.01
2026.06
67.8413.28
2026.06
67.158.9
2026.06
659.18
2026.06
659.18
2026.06
62.8617.15
2026.06
54.9215.85
2026.06
25.1832.75
2026.06
24.5241.95
2026.06
23.0660.83
2026.06
22.872.06
2026.06
21.9114.94
2026.06
21.4153.8
2026.06
2137.86
2026.06
10.2620.05
2026.06
8.214.76
2026.06
7.149.84
2026.06
6.1117.93
2026.06
5.3424.77