Artificial Analysis DeepsecBench-AA

Artificial Analysis' implementation of Vercel's DeepsecBench, testing agents on finding security vulnerabilities in real web application code. One of the three components of the AA Cyber Index.

ModelsModels with a result.21
Top resultBest result on this benchmark.46.3%GPT-6 Sol
Top-3 spreadPoints from first to third.9.6 pts
YearYear of release.2026

Result and price

19 of 21 results

Results

21 results
#ModelResultThe score from the source.
01GPT-6 Sol46.3%
05Grok 4.726.9%
10GLM-5.324.3%
11GPT-6 Luna23.5%
17Kimi K312.5%