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

Malicious Code Generation on RMCBench

92.86Attack Success Rate (ASR)

CodeSpear

-2.757622.066246.8971.7138Jun 10, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
92.8675.64
2026.06
89.9351.65
2026.06
85.5363.37
2026.06
82.7862.09
2026.06
81.6760.67
2026.06
80.9567.77
2026.06
78.2165.75
2026.06
76.3745.79
2026.06
74.1847.62
2026.06
69.656.78
2026.06
67.2225.64
2026.06
66.1221.79
2026.06
64.8423.44
2026.06
64.4746.15
2026.06
6459.33
2026.06
63.3754.76
2026.06
63.3738.83
2026.06
60.2636.45
2026.06
60.2639.74
2026.06
59.8945.6
2026.06
59.5219.23
2026.06
58.2441.58
2026.06
55.1341.03
2026.06
53.6746
2026.06
53.6640.66
2026.06
4530.67
2026.06
41.9417.58
2026.06
39.9311.36
2026.06
38.2812.45
2026.06
36.0812.82
2026.06
34.88.24
2026.06
29.857.14
2026.06
26.9223.26
2026.06
25.272.56
2026.06
23.816.41
2026.06
21.798.24
2026.06
18.324.03
2026.06
11.721.1
2026.06
107.33
2026.06
0.920.92