Print-friendly view · use your browser's Save as PDF option (Cmd/Ctrl-P) to attach this article to a brief.

Frontier-Tier AI

frontier-tier · risk class · concept

Source: https://policywindow.org/wiki/frontier-tier

Generated 2026-05-30T22:10:06 UTC

Summary

A categorical classification of AI models above certain capability or compute thresholds, indicating heightened regulatory scrutiny.

At a glance

Used by: 12 instrument(s)
Related concepts: asl-3, systemic-risk, designated-systemic, compute-threshold
Primary source: EU AI Act Art. 51 + Annex XIII (the closest binding definition)
Source URL: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1689

Details

Frontier-tier classification varies by jurisdiction. The EU AI Act presumes 'systemic risk' at ≥10²⁵ FLOPs training compute OR ≥45M EU monthly active users. The US EO 14110 used 10²⁶ FLOPs as the reporting trigger. Industry frameworks (Anthropic ASL, OpenAI Preparedness, DeepMind FSF) use capability-based rather than pure-compute frontier markers. The term 'frontier' has no single canonical definition; it is operationalized differently across regulators and developers.

How to cite this article

APA

Policy Window. (n.d.). Frontier-Tier AI [Wiki article — Concept]. https://policywindow.org/wiki/frontier-tier

Chicago

Policy Window. n.d.. "Frontier-Tier AI." Wiki article (Concept). https://policywindow.org/wiki/frontier-tier.

Harvard

Policy Window (n.d.) 'Frontier-Tier AI', Wiki article — Concept, available at: https://policywindow.org/wiki/frontier-tier.

OSCOLA

Policy Window, 'Frontier-Tier AI' (Wiki article — Concept, n.d.) <https://policywindow.org/wiki/frontier-tier> accessed [date].

BibTeX

@misc{policywindow-frontier-tier,
  title  = {Frontier-Tier AI},
  author = {Policy Window},
  year   = {n.d.},
  howpublished = {frontier-tier — risk class},
  url    = {https://policywindow.org/wiki/frontier-tier},
  note   = {Primary source: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1689}
}