Chapter VGPAI
Article 51
Classification of GPAI Models with Systemic Risk
Plain-Language Summary
A GPAI model is classified as having systemic risk if trained with more than 10^25 FLOPs, or if the Commission decides it has high-impact capabilities. This classification triggers additional obligations under Article 55.
Keywords
GPAIgeneral-purpose AIsystemic riskFLOPsfoundation modellarge language modelclassification
Legal Text
Article 51 — Classification of GPAI Models as GPAI Models with Systemic Risk 1. A general-purpose AI model shall be classified as a general-purpose AI model with systemic risk if it meets any of the following conditions: (a) it has high impact capabilities evaluated on the basis of appropriate technical tools and methodologies, pursuant to which it is determined that it has capabilities equivalent to those achievable with a cumulative amount of computation used for training measured in floating point operations (FLOPs) greater than 10^25; (b) based on a decision of the Commission, ex officio or following a qualified alert from the scientific panel, it has capabilities or an impact equivalent to those referred to in point (a). 2. A general-purpose AI model that has been classified as having systemic risk may be declassified by decision of the Commission if the provider demonstrates it does not have the relevant capabilities or impact. 3. The Commission shall, by 2 August 2025, issue guidelines specifying the practical implementation of this Article, including a list of benchmarks and tools.