We’re excited to share the technical report on EuroLLM-9B, a state-of-the-art large language model developed to serve the linguistic and cultural diversity of Europe. Trained from scratch to support all 24 official EU languages—plus 11 additional ones—EuroLLM-9B is designed to address the persistent underrepresentation of European languages in open LLMs.
This report details the model’s architecture, training data curation (including the introduction of EuroFilter), and a novel post-training dataset (EuroBlocks-Synthetic) tailored for multilingual coverage. Benchmarked against top-tier multilingual tasks, EuroLLM-9B delivers strong, competitive performance—proving that European-led, open, multilingual AI is not only possible but leading the way.
🌍 Want to explore how EuroLLM-9B is setting a new standard for open multilingual models?
Download the full report and see how this breakthrough model is shaping the future of inclusive, transparent AI in Europe.
