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

Multi-task Language Understanding on MMLU (Accuracy)

74.4Accuracy

SA-SFT

20.8434.74548.6562.555Nov 25, 2024Feb 8, 2025Apr 24, 2025Jul 8, 2025Sep 21, 2025Dec 5, 2025Feb 19, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.01
74.4
2026.02
74.3
2026.02
73.73
2026.02
72.92
2026.02
72.82
2026.02
72.52
2026.02
72.49
2026.02
70.41
2026.02
70.36
2026.02
70.35
2026.02
70.27
2026.02
70.18
2026.02
70.15
2026.02
70.08
2026.02
69.94
2026.02
69.9
2026.02
69.65
2025.12
68.36
2025.12
68.32
2025.12
68.3
2026.02
68.25
2025.12
68.2
2025.12
68.14
2025.12
68.1
2025.12
68.05
2025.12
68
2025.12
67.82
2025.12
67.74
2025.12
67.05
2026.01
65.3
2026.02
63.3
2025.09
62.2
2025.12
62.1
2026.01
59.8
2026.02
59.67
2026.02
59.61
2026.02
59.47
2026.02
59.36
2026.02
59.32
2026.01
59.2
2026.02
56.77
2026.02
56.35
2026.02
56.28
2026.02
56.28
2026.02
55.89
2026.01
53.7
2026.02
52.14
2025.09
51.7
2026.02
50.38
2025.12
48
2024.12
47
2024.12
46.8
2024.12
45.5
2025.12
44.8
2024.12
44.6
2024.12
43.1
2025.12
42.9
2025.09
42.8
2025.12
41.7
2025.12
41.3
2025.12
37.8
2025.12
35.3
2025.12
34
2025.12
33.2
2025.12
32.9
2025.09
32.2
2025.12
31.4
2025.09
30.6
2025.12
29.4
2025.12
29.3
2024.12
28.8
2025.12
28.5
2025.12
26.7
2024.12
26.5
2025.09
26.4
2025.09
25.4
2024.11
25.35
2024.12
25.3
2024.11
25.29
2025.12
25.1
2025.09
25
2025.12
24.8
2025.12
24.8
2026.02
24.78
2025.09
23.2
2025.12
23
2025.09
22.9