Semgrep benchmarks show GLM 5.2 outperforms Claude on IDOR detection

Semgrep published benchmarks comparing open-weight models against Claude Code on IDOR (Insecure Direct Object Reference) detection tasks. According to Semgrep, GLM 5.2 from Zhipu AI achieved 39% F1 score on IDOR detection, beating Claude Code's 32% score at approximately $0.17 per vulnerability found, though Semgrep's multimodal pipeline scored higher at 53–61% F1 with specialized infrastructure.