AI Security Leaderboard
Compare AI model performance across authentication, authorization, and vulnerability detection challenges
47
Total Models
AI models evaluated
1,247
Security Tests
Completed evaluations
78.3%
Average Score
Across all categories
2 hours
Last Updated
ago
Model Rankings
AI models ranked by overall security performance
GPT-4 Turbo
VerifiedOpenAI
1247 tests • 2 hours ago
94.2%
Overall
96.1%
Auth
92.8%
AuthZ
93.7%
Vuln
+2.1%
Claude-3 Opus
VerifiedAnthropic
1198 tests • 4 hours ago
91.8%
Overall
93.2%
Auth
89.1%
AuthZ
92.9%
Vuln
+0.7%
Gemini Ultra
Verified1156 tests • 6 hours ago
89.4%
Overall
91.7%
Auth
86.3%
AuthZ
90.2%
Vuln
-1.2%
#4
Claude-3 Sonnet
VerifiedAnthropic
1089 tests • 8 hours ago
87.6%
Overall
89.4%
Auth
84.8%
AuthZ
88.6%
Vuln
+1.8%
#5
GPT-4
VerifiedOpenAI
1234 tests • 12 hours ago
85.3%
Overall
87.1%
Auth
82.9%
AuthZ
86%
Vuln
0%
#6
Llama-2 70B
Meta
967 tests • 1 day ago
82.7%
Overall
84.2%
Auth
79.8%
AuthZ
84.1%
Vuln
+3.2%
#7
Gemini Pro
Verified1045 tests • 1 day ago
81.4%
Overall
83.6%
Auth
78.1%
AuthZ
82.5%
Vuln
-0.9%
#8
Claude-3 Haiku
VerifiedAnthropic
892 tests • 2 days ago
79.8%
Overall
81.3%
Auth
76.9%
AuthZ
81.2%
Vuln
+1.4%