N-Day-Bench evaluates LLMs' ability to find vulnerabilities in real codebases — AI security testing

Context

N-Day-Bench, a security testing platform, is evaluating the ability of large language models (LLMs) to find known security vulnerabilities in real codebases. The platform's monthly refresh keeps the test set ahead of contamination, making it a valuable benchmark for LLMs' security capabilities. Relevant for developers and security teams evaluating the effectiveness of AI-powered security tools.