Artificial Analysis CWE-Bench-AA

Artificial Analysis' implementation of Collinear AI's CWE-bench, testing whether agents can patch real-world software vulnerabilities. One of the three components of the AA Cyber Index.

ModelsModels with a result.19
Top resultBest result on this benchmark.68.3%Grok 4.7
Top-3 spreadPoints from first to third.4.1 pts
YearYear of release.2026

Result and price

17 of 19 results

Results

19 results
#ModelResultThe score from the source.
01Grok 4.768.3%
03GPT-6 Sol64.2%
09GPT-6 Luna56.7%
10GLM-5.355.8%
12Kimi K353.3%
17MiniMax M341.7%