Security/ security · ai · benchmarking

LLM Agents Solve Half of Real CVEs, and the Pricier Model Doesn't Win

A 300-run benchmark on 20 real CVEs found LLM agents top out at a 50% fix rate, and the most expensive model costs 12 times more for no gain.

A new benchmark pits LLM agents against 20 real security vulnerabilities, and the best performer fixes only half.

The project, CVE-Bench, drew its test cases from 18 Python projects including Pillow, GitPython, yt-dlp, and urllib3. Five agents — three from OpenAI and two from poolside — were each assigned a vulnerability to patch inside a sandbox, then scored against security tests drawn from the maintainers' own fixes. Across 300 runs and three prompt styles (full advisory, locate, and diagnose), the top solve rate was 50%. The remaining half is the interesting part: some of those failed fixes pass every regression test while leaving the vulnerability intact.

That false-pass result isn't a footnote. A tool that turns CI green while a CVE survives doesn't just fail — it erases the signal that something is wrong, which may be harder to recover from than an obvious miss. The cost picture adds another reason for skepticism: gpt-5.5 ran at 12 times the price of gpt-5.4-mini and produced statistically similar results, suggesting the premium-model reflex in security tooling isn't backed up here.

The author's power analysis puts the task count needed to detect a meaningful within-family performance gap at roughly 700 — most published LLM security benchmarks don't get close, so their leaderboards may be telling us less than they appear to.

TR

The Revision

Written by an AI system from the public sources credited above. How we write →