The Frontier Model Forum is an industry non-profit dedicated to advancing the safe development and deployment of frontier AI systems. 

The FMF draws on the technical and operational expertise of its member firms to benefit the broader AI safety ecosystem. We seek to admit members with a clear ability to develop and deploy frontier AI systems at scale, a proven commitment to AI safety, and a willingness to contribute to the mission of the FMF.  

Member criteria

The FMF has three high level criteria:

Frontier AI capability

  • Proven ability: Member firms must provide evidence that they have the expertise and resources to develop or deploy at scale frontier AI models.

Track record of safety

  • Documentation: Member firms must publicly acknowledge that frontier AI models pose both public safety and societal risks, and publicly disclose guidelines for evaluating and mitigating those risks. 
  • Safety review processes: Member firms must provide evidence that they have processes in place for assessing the safety of frontier models throughout their lifecycle, and that those processes may lead to the delay or suspension of model deployment. 
  • Research and evaluations: Member firms must provide evidence that they support AI safety research, including both third-party evaluations and independent studies on the societal impact of frontier AI. 

Clear intentions to contribute: 

  • Commitment to the FMF: Member firms must commit to participating in FMF workshops and activities, and to financially supporting the FMF’s safety efforts for a minimum of three years. 

Contact us

To inquire about FMF membership, please contact