Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Prompt Injection on URL-based PI (200-sample dataset)
Loading...
33.5
ASR
Standard
-1.34
7.705
16.75
25.795
Feb 5, 2026
ASR
Updated 4d ago
Evaluation Results
Method
Method
Links
ASR
Standard
Model=Grok 4 Fast
2026.02
33.5
Standard
Model=Llama 4
2026.02
22
Standard
Model=GPT-5
2026.02
20
Advanced
Model=Llama 4
2026.02
16
InjectDefuser
Model=Grok 4 Fast
2026.02
8
InjectDefuser
Model=Llama 4
2026.02
7
Advanced
Model=Grok 4 Fast
2026.02
2
Advanced
Model=GPT-5
2026.02
1
Standard
Model=Gemma 3
2026.02
0
Advanced
Model=Gemma 3
2026.02
0
InjectDefuser
Model=GPT-5
2026.02
0
InjectDefuser
Model=Gemma 3
2026.02
0
Feedback
Search any
task
Search any
task