Skip to main content
TipRanks
Go Ultimate75% OFF

Early Benchmark Compares Evo Code Scanning Agent With AI Security Tools

Early Benchmark Compares Evo Code Scanning Agent With AI Security Tools

According to a recent LinkedIn post from Snyk, an internal benchmark compared the security harness in its Evo Code Scanning Agent with systems using the same models associated with Claude Security and Codex Security. The evaluation covered 14 applications and 1,426 known vulnerabilities under the same model and compute conditions.

The post reports detection comparisons of 57% versus 32% on Claude Sonnet 5.5 and 66% versus 43% on GPT-6.1 Sol. Snyk characterizes the findings as early results from an experimental preview, suggesting that performance depends not only on the underlying model but also on the security tooling and workflow surrounding it.

If replicated in broader, independent testing, the results could support Snyk’s positioning in agentic application security and strengthen the case for Evo Code Scanning Agent within its Evo Agentic AppSec offering. However, the limited sample, internal methodology and experimental status make the figures directional rather than evidence of established commercial or financial impact.

Disclaimer & DisclosureReport an Issue

Related Articles