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

Code Generation on MBPP (Performance %)

83Performance (%)

DuoAttention

-3.3219.0941.563.91Oct 9, 2025
Updated 6d ago

Evaluation Results

MethodLinks
2025.10
83
2025.10
82.4
2025.10
81
2025.10
78.2
2025.10
76.4
2025.10
74.4
2025.10
67
2025.10
63.8
2025.10
63.8
2025.10
63.2
2025.10
62.8
2025.10
62
2025.10
60.6
2025.10
60.4
2025.10
59.8
2025.10
57.8
2025.10
56.8
2025.10
56.6
2025.10
55.2
2025.10
55
2025.10
51.8
2025.10
50
2025.10
46.6
2025.10
46.4
2025.10
41.4
2025.10
40.4
2025.10
40.2
2025.10
39.8
2025.10
39.6
2025.10
39.2
2025.10
35
2025.10
31
2025.10
29.4
2025.10
14.6
2025.10
12.6
2025.10
12.6
2025.10
12.2
2025.10
10.8
2025.10
7.4
2025.10
7.2
2025.10
6
2025.10
5.2
2025.10
1.6
2025.10
1
2025.10
1
2025.10
1
2025.10
1
2025.10
0.6
2025.10
0.6
2025.10
0.4
2025.10
0.4
2025.10
0.4
2025.10
0.4
2025.10
0.2
2025.10
0.2
2025.10
0
2025.10
0
2025.10
0
2025.10
0
2025.10
0