Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Refusal Evaluation on Harmful prompt suite (test)
Loading...
95.5
Refusal Rate
Base
-3.82
21.965
47.75
73.535
Jan 13, 2026
Refusal Rate
Updated 4d ago
Evaluation Results
Method
Method
Links
Refusal Rate
Base
Model=Ministral-3B-Ins...
2026.01
95.5
Base
Model=Qwen3-VL-8B-Inst...
2026.01
93.8
Base
Model=Ministral-14B-In...
2026.01
91.9
Base
Model=Qwen3-VL-4B-Inst...
2026.01
84
Base
Model=Qwen3-VL-2B-Inst...
2026.01
83.3
Standard
Model=Qwen3-VL-8B-Inst...
2026.01
42
Standard
Model=Ministral-14B-In...
2026.01
12
SRA
Model=Qwen3-VL-8B-Inst...
2026.01
2
SRA
Model=Ministral-3B-Ins...
2026.01
2
Standard
Model=Qwen3-VL-2B-Inst...
2026.01
0
SRA
Model=Qwen3-VL-2B-Inst...
2026.01
0
Standard
Model=Qwen3-VL-4B-Inst...
2026.01
0
SRA
Model=Qwen3-VL-4B-Inst...
2026.01
0
Standard
Model=Ministral-3B-Ins...
2026.01
0
SRA
Model=Ministral-14B-In...
2026.01
0
Feedback
Search any
task
Search any
task