# OpenAI and the Frontier Model Forum launch a jailbreak-harm rating framework



OpenAI



OpenAI and Frontier Model Forum member firms launched a framework that rates when bypassing an AI system's safeguards could enable serious cyber harm, so protections can be prioritized.

- Recorded time: Nov 07, 2026 08:31 · Beijing Time (UTC+8)
- Review updated: Nov 07, 2026 11:29 Sat · Beijing Time (UTC+8)

## Sources

- Source excerpt: https://x.com/OpenAI/status/2118795218099253606

Canonical entry: https://www.ailyre.com/source/?id=openai-fmf-jailbreak-rating-framework-2026-11-06&lang=en