The OpenEuroLLM Project, an initiative designed to create open-source artificial intelligence (AI) models, was officially launched on Monday. The project has garnered support from the European Commission, which has recognized its significance by awarding it the Strategic Technologies for Europe Platform (STEP) Seal, signifying its status as a key technology endeavor. The primary goal of this alliance is to develop a range of multilingual large language models (LLMs) capable of operating effectively across all languages used within the European Union (EU). Transparency and compliance with regulatory standards established by the bloc will guide the development process.
European Commission Supports the OpenEuroLLM Initiative
In a statement published on X (previously known as Twitter), the European Commission’s official account announced that the OpenEuroLLM Project is the inaugural recipient of the STEP Seal this year. This quality label is assigned to premium projects that meet established quality benchmarks under the Digital Europe Programme. Selected projects benefit from increased visibility on the STEP platform, aimed at attracting potential investors.
According to a press release from the OpenEuroLLM Project, work on developing the AI models commenced on February 1. The initiative is a collaboration involving a consortium of 20 European research institutions, companies, and EuroHPC centres. The project is led by Jan Hajič from Charles University in Czechia, alongside Peter Sarlin, Co-Founder and CVP at AMD Silo AI.
The OpenEuroLLM Project envisions the creation of a suite of high-performance multilingual LLMs that will be made available to the open-source community for applications in commercial, industrial, and public services. The project has committed to upholding stringent regulatory policies set forth by the EU and aims to maintain transparency in its data procurement processes.
Upon the release of the AI models, the project will also provide comprehensive documentation, including training and testing code, as well as evaluation metrics for the models. This initiative aims to enable fine-tuning and instruction-tuning to meet the specific needs of various industries and the public sector.
According to the press release, “The transparent and compliant open-source models will democratize access to high-quality AI technologies and enhance the capability of European companies to compete globally, while also empowering public organizations to deliver impactful services.”
The European Commission has already allocated funding for the OpenEuroLLM Project through the Digital Europe Programme, with expectations of attracting further investors in the near future. Currently, there is no timeline available for the release of these models, nor has any specific area of focus for their development been disclosed.
