Vercel Launches DeepsecBench to Rank AI Models on Cybersecurity Vulnerability Detection

Vercel launches DeepsecBench, a new benchmark ranking AI models on cybersecurity vulnerability detection, revealing that while GPT-5.6 Sol tops the leaderboard, budget-friendly models like Grok 4.5 and Kimi K3 offer competitive security scanning at a fraction of the cost.

• Vercel releases DeepsecBench, a benchmark that evaluates how well AI models detect cybersecurity vulnerabilities in application code, scoring models on recall, precision, cost, and total scan time.
• GPT-5.6 Sol leads the leaderboard with a score of 35.58 at $55.98, while cost-efficient alternatives like Grok 4.5 and Kimi K3 deliver competitive performance at a fraction of the price, signaling that capable security scanning is becoming more affordable.
• Organizations can now build multi-model scanning programs using deepsec and AI Gateway, running frontier models for deep periodic audits while deploying cheaper models for continuous or pre-deployment scans across their entire codebase.