Dataset

Public AI Security Index

Security teams are increasingly tasked with approving AI models and MCP servers despite a critical lack of standardized security data. The Public AI Security Index provides independent, verifiable benchmarks for LLMs and MCP servers, powered by AQtive Guard AI‑SPM, so you can make defensible architecture decisions.

Quickly identify which AI components meet your security baseline and which require compensating controls.

Download the full .json dataset to:

  • Access ranked security scores for popular LLMs, including insights into jailbreak susceptibility, misuse potential, toxicity, and robustness.
  • Evaluate vulnerability and severity metrics for widely deployed MCP servers.
  • Integrate seamlessly: Plug these raw benchmarks directly into your existing risk models, SIEMs, and custom dashboards.

Download Dataset