SecCodeBench

SecCodeBench evaluates LLM coding agents on secure code generation and vulnerability detection, testing the ability to produce code that is both functional and free from security vulnerabilities.

id seccodebench · max 1 · 2 models reported