OpenEuroLLM: Open, Multilingual, European LLMs

OpenEuroLLM is a collaborative initiative dedicated to building open, multilingual, and foundation Large Language Models (LLMs) tailored for the European context. Our mission is to advance generative AI technology while ensuring transparency, accessibility, and linguistic diversity.

OpenEuroLLM is a project co-funded by the Digital Europe Programme under GA no. 101195233. For more information see openeurollm.eu.

🧠 Core Interests

We focus on the intersection of:

• Open Science: Developing models and datasets that are publicly accessible.
• Multilingualism: Supporting European languages and cross-lingual capabilities.
• Transparent AI: Pushing the boundaries of open foundation models.
• European Ecosystem: Fostering sovereign AI capabilities within Europe.

🤝 Get Involved

We believe in the power of community-driven development. Whether you are a researcher, developer, or enthusiast, your contributions are very welcome:

• Explore our models and datasets.
• Report issues or suggest improvements.

🔗 Links

• Website: openeurollm
• Our github: https://github.com/OpenEuroLLM/
• Our task board: where we discuss our progress and new results as they come https://github.com/OpenEuroLLM/Taskboard
• Training code: the library we use to manage our training runs at scale https://github.com/OpenEuroLLM/oellm-autoexp/
• Evaluation code: the library we use to schedule large evaluation campaigns on several EuroHPC machines for many eval frameworks (lm-harness, light-eval, evalchemy, judgearena / LLM judges) https://github.com/OpenEuroLLM/oellm-evals/
• Papers: papers published from the project (bibtex)

Building the future of European AI, openly.