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

Jailbreak Success Rate (GPT-4o/5 Judges) and Query Cost on AdvBench

99.1ASR (GPT-4o)

HMNS

37.42853.43969.4585.461Apr 11, 2026
Updated 5d ago

Evaluation Results

MethodLinks
2026.04
99.195.21.8
2026.04
97.492.52
2026.04
93.489.27.3
2026.04
92.986.82
2026.04
91.186.27.6
2026.04
86.880.78.3
2026.04
73.967.612.3
2026.04
70.864.612.7
2026.04
64.958.213.8
2026.04
45.940.115.6
2026.04
4236.516.4
2026.04
39.83417.1