Article

TildeOpen, an open-source foundational model for European languages, was released on the Hugging Face platform in autumn 2025. Developed by Tilde for the European Commission, the model contains 30 billion parameters and is trained on hundreds of billions of words in various European languages, with 29 billion text units from Latvian.

This marks the largest dataset utilized in developing Latvian artificial intelligence. The model's creation followed Tilde's victory in the Large AI Grand Challenge contest, and it was developed using the LUMI supercomputer in Finland.