Veracode found that the average security pass rate for AI-generated code sits at roughly 56% across more than 100 models and four snapshots, with 44% of generation tasks introducing detectable known vulnerabilities.
Public source
Publisher name
Public post
#AI is getting better at writing code that works. Secure code is still a separate challenge. That is one of the key points in Chris Wysopal’s latest blog on OpenAI’s cyb…
Company
Veracode
Transforming application risk management for the AI era.
- Industry
- Computer and Network Security
- Location
- Burlington, US
- Company size
- 501–1,000 employees
About Veracode
Veracode is a global leader in Application Risk Management for the AI era. Powered by trillions of lines of code scans and a proprietary AI-assisted remediation engine, the Veracode platform offers adaptive software security and is trusted by organizations worldwide to build and maintain secure software from code creation to cloud deployment. Thousands of the world’s leading development and security teams use Veracode every second of every day to get accurate, actionable visibility of exploitable risk, achieve real-time vulnerability remediation, and reduce their security debt at scale. Veracode is a multi-award-winning company offering capabilities to secure the entire software development life cycle, including Veracode Fix, Static Analysis, Dynamic Analysis, Software Composition Analysis, Container Security, Application Security Posture Management, Malicious Package Detection, and Penetration Testing.
See moreLatest activity
Latest activity from Veracode
30 signals
Research & Knowledge
Veracode research shows that roughly 40 to 45 percent of AI-generated code contains security vulnerabilities.
Presence & Recognition
Veracode is exhibiting at GISEC GLOBAL alongside Green Method Technologies at Booth B180 in Dubai from September 16 to 18, 2026.
Presence & Recognition
Veracode will speak at the Billington CyberSecurity Summit at Breakout C6 with Chris Wysopal.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
Yubico
Yubico tested top AI models across 29 codebases to see what works in practice and found that AI delivered 448 verified security hygiene fixes, human engineers downgraded 46% of overhyped severity ratings, and multi-stage prompts proved crucial for accuracy.
Research & Knowledge
Contrast Security
Contrast Security set three AI scanners loose on the same codebase and found they agreed on only 5 percent of the findings.
Research & Knowledge
ProjectDiscovery
ProjectDiscovery conducted research attributing AI assistance to the gap between AI-generated code and security capacity.
Research & Knowledge
Secure Code Warrior
Secure Code Warrior published an in-depth study analyzing 16 leading AI models across 11 real-world frameworks and 1,760 codebases to uncover insights on AI coding security.
Research & Knowledge
Radware