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

Massive Multitask Language Understanding on MMLU (Overall Accuracy)

84.2MMLU Accuracy

MARI

20.65637.15353.6570.147Nov 14, 2025Dec 16, 2025Jan 17, 2026Feb 19, 2026Mar 23, 2026Apr 24, 2026May 27, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.05
84.2
2026.05
83.7
2026.05
83.6
2026.05
83.5
2026.05
83.4
2026.05
83.2
2026.05
81.6
2026.05
81.2
2026.05
81
2026.05
80.9
2026.05
80.8
2026.05
80.6
2025.11
76.6
2026.05
76.58
2026.05
76.58
2026.05
76.34
2026.05
76.25
2026.05
76.08
2026.05
76.08
2026.05
76
2026.05
73.1
2026.05
72.6
2026.05
72.5
2026.05
72.4
2026.05
72.3
2026.05
72.1
2025.11
71.3
2025.11
69.1
2025.11
68.5
2025.11
68.2
2026.05
66.6
2026.05
66.1
2026.05
66
2026.05
65.9
2026.05
65.8
2026.05
65.5
2025.11
63.9
2025.11
62
2025.11
56.5
2025.11
51.6
2025.11
49.5
2025.11
48.9
2025.11
46.3
2026.05
26.7
2026.05
26.3
2026.05
26.1
2026.05
26
2026.05
25.9
2026.05
25.7
2026.05
23.6
2026.05
23.4
2026.05
23.3
2026.05
23.2
2026.05
23.2
2026.05
23.1