List your company

Jailbreak on HarmBench

1.01Toxicity ScoreFigStep
0.85921.87712.8953.9129May 7, 2026

Evaluation results

MethodLinks
FigStepVictim Model=GPT-5.4-mini2026.051.010.610.270.53——
HADESVictim Model=Claude Haiku 4.52026.051.010.060.671.47——
FlipAttackVictim Model=Claude Haiku 4.52026.051.010.110.531.07——
HADESVictim Model=GPT-5.4-mini2026.051.020.490.671.73——
FigStepVictim Model=Claude Haiku 4.52026.051.030.080.531.33——
FlipAttackVictim Model=GPT-5.4-mini2026.051.040.660.531.87——
HADESVictim Model=GPT-5.4-nano2026.051.050.881.331.73——
SIVictim Model=GPT-5.4-nano2026.051.050.891.21.92——
SIVictim Model=Claude Haiku 4.52026.051.050.21.281.84——
SIVictim Model=GPT-5.4-mini2026.051.060.541.842.32——
FigStepVictim Model=GPT-5.4-nano2026.051.111.233.23.33——
FlipAttackVictim Model=GPT-5.4-nano2026.051.160.724.85.6——
TxtV-GDIVictim Model=Claude Haiku 4.52026.051.220.545.796.45——
CS-DJVictim Model=Claude Haiku 4.52026.051.280.86.85.73——
CS-DJVictim Model=GPT-5.4-mini2026.051.41.8311.27.6——
TxtVVictim Model=Claude Haiku 4.52026.051.40.9910.5910.11——
TTV-GDIVictim Model=Claude Haiku 4.52026.051.541.2814.4814.13——
TxtVVictim Model=GPT-5.4-nano2026.051.611.7816.2411.25——
TypVVictim Model=Claude Haiku 4.52026.051.651.4817.618.21——
TTVVictim Model=Claude Haiku 4.52026.051.651.5217.1517.12——
CS-DJVictim Model=GPT-5.4-nano2026.051.862.7525.7314.4——
SIVictim Model=Gemini-2.5-Flash2026.051.922.0123.0723.31——
TxtV-GDIVictim Model=GPT-5.4-mini2026.051.982.2424.9624.37——
TypVVictim Model=GPT-5.4-mini2026.051.992.6326.1124.88——
TypVVictim Model=GPT-5.4-nano2026.0522.4926.2721.18——
TxtVVictim Model=GPT-5.4-mini2026.052.162.7430.7229.06——
HADESVictim Model=Gemini-2.5-Flash2026.052.182.6831.3327.33——
TxtV-GDIVictim Model=GPT-5.4-nano2026.052.283.1433.5725.76——
TTVVictim Model=GPT-5.4-mini2026.052.313.0134.2633.12——
TTVVictim Model=GPT-5.4-nano2026.052.333.1834.9630.06——
TTV-GDIVictim Model=GPT-5.4-nano2026.052.553.7240.2934.21——
TTV-GDIVictim Model=GPT-5.4-mini2026.052.613.4741.9441.41——
HADESVictim Model=Gemini-3.1-Flash-Lite2026.052.753.9645.8742.13——
FigStepVictim Model=Gemini-3.1-Flash-Lite2026.052.924.1550.1349.2——
SIVictim Model=Gemini-3.1-Flash-Lite2026.053.144.8755.8143.65——
CS-DJVictim Model=Gemini-3.1-Flash-Lite2026.053.465.764.2737.07——
CS-DJVictim Model=Gemini-2.5-Flash2026.053.716.1475.244.93——
TTVVictim Model=Gemini-2.5-Flash2026.054.036.8778.1970.72——
TypVVictim Model=Gemini-3.1-Flash-Lite2026.054.127.1280.5977.89——
TypVVictim Model=Gemini-2.5-Flash2026.054.137.0180.2474.96——
TTVVictim Model=Gemini-3.1-Flash-Lite2026.054.257.3182.9181.06——
TTV-GDIVictim Model=Gemini-2.5-Flash2026.054.277.3584.0375.65——
TxtV-GDIVictim Model=Gemini-2.5-Flash2026.054.347.7185.9574.56——
TTV-GDIVictim Model=Gemini-3.1-Flash-Lite2026.054.347.5185.7383.01——
TxtVVictim Model=Gemini-2.5-Flash2026.054.367.6485.8780.77——
TxtV-GDIVictim Model=Gemini-3.1-Flash-Lite2026.054.447.6688.4885.55——
FigStepVictim Model=Gemini-2.5-Flash2026.054.497.7189.288.27——
TxtVVictim Model=Gemini-3.1-Flash-Lite2026.054.587.9991.3389.47——
FlipAttackVictim Model=Gemini-3.1-Flash-Lite2026.054.728.5294.5393.2——
FlipAttackVictim Model=Gemini-2.5-Flash2026.054.788.7995.0783.07——
AGRTarget Model=Qwen3-1.7B2026.05————9694
AGRTarget Model=Qwen3-8B2026.05————9696.7
AGRTarget Model=DS-R1-Distill-Llama-8B2026.05————9897.3
AutoDANTarget Model=Qwen3-1.7B2026.05————42.741.3
AutoDANTarget Model=Qwen3-8B2026.05————37.336
AutoDANTarget Model=DS-R1-Distill-Llama-8B2026.05————33.332.7
AutoRANTarget Model=Qwen3-1.7B2026.05————90.789.3
AutoRANTarget Model=Qwen3-8B2026.05————87.388
AutoRANTarget Model=DS-R1-Distill-Llama-8B2026.05————90.790.7
GCGTarget Model=Qwen3-1.7B2026.05————3230.7
GCGTarget Model=Qwen3-8B2026.05————20.719.3
GCGTarget Model=DS-R1-Distill-Llama-8B2026.05————3030
H-CoTTarget Model=Qwen3-1.7B2026.05————9090.7
H-CoTTarget Model=Qwen3-8B2026.05————9090.7
H-CoTTarget Model=DS-R1-Distill-Llama-8B2026.05————89.390
PAIRTarget Model=Qwen3-1.7B2026.05————3433.3
PAIRTarget Model=Qwen3-8B2026.05————29.328.7
PAIRTarget Model=DS-R1-Distill-Llama-8B2026.05————22.722.7
ReNeLLMTarget Model=Qwen3-1.7B2026.05————6060
ReNeLLMTarget Model=Qwen3-8B2026.05————5252.7
ReNeLLMTarget Model=DS-R1-Distill-Llama-8B2026.05————67.368