Skip to main content
XOR

Double Free

The product calls free() twice on the same memory address.

CWE-415 at MITRE →

53.5%

repair rate

0–80

95% interval

43

scored attempts · 3 codebases

By model

#AgentModelRepair rate with intervalRateAttempts
1gemini31gemini-3.1-pro-preview100.0% — no interval: too few codebases or attempts to generalise100.0%2
2claudeclaude-opus-4-666.7% — no interval: too few codebases or attempts to generalise66.7%3
3codexgpt-5.266.7% — no interval: too few codebases or attempts to generalise66.7%3
4cursorgpt-5.266.7% — no interval: too few codebases or attempts to generalise66.7%3
5cursorgpt-5.3-codex66.7% — no interval: too few codebases or attempts to generalise66.7%3
6cursoropus-4.666.7% — no interval: too few codebases or attempts to generalise66.7%3
7opencodeclaude-opus-4-666.7% — no interval: too few codebases or attempts to generalise66.7%3
8opencodegpt-5.266.7% — no interval: too few codebases or attempts to generalise66.7%3
9opencodegpt-5.2-codex50.0% — no interval: too few codebases or attempts to generalise50.0%2
10claudeclaude-opus-4-533.3% — no interval: too few codebases or attempts to generalise33.3%3
11codexgpt-5.2-codex33.3% — no interval: too few codebases or attempts to generalise33.3%3
12cursorcomposer-1.533.3% — no interval: too few codebases or attempts to generalise33.3%3
13geminigemini-3-pro-preview33.3% — no interval: too few codebases or attempts to generalise33.3%3
14opencodeclaude-opus-4-533.3% — no interval: too few codebases or attempts to generalise33.3%3
15opencodegemini-3.1-pro-preview33.3% — no interval: too few codebases or attempts to generalise33.3%3

outcome mix

23 repaired · 4 not repaired · 16 did not build

23 repaired · 4 not repaired · 16 did not build

Corpus cve-bench-136 · run February 2026 · How we test

All weakness classes →