V
83.0%
Vulnaut
AuditAgent Agent
EVMBench public audit comparison
Detection Rate Comparison
Vulnerability detection rate (%) on the EVMBench dataset. Higher is better.
Lead score
83.0%
Gap to next
+7.8
Systems compared
16
D
75.2%
Azimuth Agent
TestMachine
N
67.0%
AuditAgentAgent
Nethermind Security
G
59.8%
Guardix Agent
Guardix
A
45.6%
Claude Opus 4.6
Anthropic
O
39.2%
GPT-5.3-Codex
xhigh reasoning
O
39.2%
GPT-5.2
xhigh reasoning
A
36.1%
Claude Opus 4.5
Anthropic
O
34.2%
GPT-5.3-Codex
high reasoning
O
30.0%
OC-GPT-5.2
OpenAI
O
29.7%
GPT-5.2
high reasoning
O
29.7%
GPT-5.2
medium reasoning
O
26.9%
GPT-5.3-Codex
medium reasoning
O
23.3%
GPT-5
OpenAI
O
22.2%
GPT-5.2
low reasoning
G
20.8%
Gemini 3 Pro
Google