# Manage systemic risk for high-impact general-purpose models

- **Record type**: Obligation
- **Instrument**: Regulation (EU) 2024/1689 — Artificial Intelligence Act
- **Jurisdiction**: European Union
- **Category**: safety_testing
- **Binding**: Yes
- **Applies from**: 2025-08-02
- **Provision**: Articles 51, 52 and 55

## In brief

Manage systemic risk for high-impact general-purpose models is a legal requirement under the EU AI Act in the European Union (Articles 51, 52 and 55). A general-purpose model is presumed to have systemic risk when the cumulative compute used for training exceeds 10^25 floating-point operations, or when the Commission designates it. It applies from 2 August 2025.

## What the duty requires

A general-purpose model is presumed to have systemic risk when the cumulative compute used for training exceeds 10^25 floating-point operations, or when the Commission designates it. Providers must notify the Commission, perform model evaluations including adversarial testing, assess and mitigate systemic risks, track and report serious incidents, and ensure adequate cybersecurity for the model and infrastructure.

## What an organisation does about it

Track training compute against the threshold and prepare a safety and security framework before launch.

## Controls that meet this duty

- Frontier model safety and security framework (satisfies): https://aipolicytracker.org/controls/frontier-model-safety-framework
- Adversarial and red-team testing for generative AI (satisfies): https://aipolicytracker.org/controls/genai-red-team-testing
- AI incident management and regulatory reporting (supports): https://aipolicytracker.org/controls/ai-incident-management-and-reporting

## Parent instrument

- Regulation (EU) 2024/1689 — Artificial Intelligence Act
  - Record: https://aipolicytracker.org/policies/eu-ai-act
  - Context file: https://aipolicytracker.org/policies/eu-ai-act.md

## Provenance

- **Permanent id**: AIPT-OBL-0019 (https://aipolicytracker.org/id/AIPT-OBL-0019; never changes or gets reused)
- **Record page**: https://aipolicytracker.org/obligations/eu-ai-act-gpai-systemic-risk
- **Official source**: https://eur-lex.europa.eu/eli/reg/2024/1689/oj
- **Provision**: Articles 51, 52 and 55
- **Review status**: pending review
- **Confidence**: medium
- **Facts last confirmed**: never confirmed against the official source
- **Retrieved**: 2026-10-27
- **Licence**: https://creativecommons.org/licenses/by/4.0/
- **Cite as**: AIPolicyTracker (2026). "Manage systemic risk for high-impact general-purpose models (EU AI Act)" [AIPT-OBL-0019, not yet verified]. https://aipolicytracker.org/id/AIPT-OBL-0019 (accessed 27 October 2026).

> This record is a structured summary with a link to the official text. It is not legal advice. Open the official source before relying on any date or duty. How current each record type must be is published at https://aipolicytracker.org/verification; what a record must carry at all is published at https://aipolicytracker.org/coverage.