GLM 5.2 Beats Claude Opus 4.8 on Semgrep's Real-World IDOR Test
Semgrep's IDOR benchmark pitted GLM 5.2 against Claude Opus 4.8 with identical prompts. The open-weight model won — and that changes the cybersecurity math.
Semgrep's IDOR benchmark pitted GLM 5.2 against Claude Opus 4.8 with identical prompts. The open-weight model won — and that changes the cybersecurity math.