Skip to main content
Rase AI Intelligence Newsletter

Models & assistants Model or AI service

EuroLLM-9B covers the European Union’s 24 official languages

The EuroLLM team released a 9B multilingual language model with base and instruction-tuned checkpoints. It trained the model for the European Union’s 24 official languages and 11 others. [1][2]

Model access
Downloadable base and instruction-tuned 9B checkpoints. The repository lists access conditions. [1][2]
License
Apache License 2.0 for the documented 9B weights. [2]

What the model does

The project introduced EuroLLM-9B in December 2024, following its earlier 1.7B model. Both 9B versions were linked in the release announcement. [1][2]

Why it matters

A team handling translation or multilingual documents can evaluate a single model across the languages it serves. The project’s public checkpoints make it possible to repeat the same assessment without requiring a fixed hosted service. [1][2]

Who can use it

Researchers can download the base checkpoint. Developers can also evaluate the instruction-tuned version, which the project tuned for instruction following and translation. [1][2]

Key dates

  1. EuroLLM-9B base and Instruct released [1]

Ways to use it

Compare translations produced by the instruction-tuned version with human reference translations in your target language. Test specialist terminology as well as general language. [1][2]

Model limitations

Language coverage does not establish equal accuracy across all languages. The published repository may ask users to agree to its access conditions before it provides the files. [1][2]

Sources and verification

Catalog evidence checked How we verify records

  1. Primary source EuroLLM team’s December 2 release of EuroLLM-9B
  2. Primary source utter-project/EuroLLM-9B-Instruct
  3. Independent source EuroLLM-9B Aims to Improve State of the Art LLM Support for European Languages - InfoQ
  4. Primary source Hugging Face registry metadata for utter-project/EuroLLM-9B-Instruct

Related profiles

Models & assistants

GPT-6.2 Astra and Sol

OpenAI

Astra targets demanding projects. Sol targets everyday tasks. Both support text and image input in the API.

High confidenceEvidence checked

Models & assistants

MAI Code 2 and Code 2 Flash

Microsoft AI

Code 2 targets complex software work. Flash targets frequent edits, explanations and routine fixes.

High confidenceEvidence checked

Models & assistants

MAI Realtime 1

Microsoft AI

Users can try a MAI Labs demo. Developers can register for a limited Foundry preview.

High confidenceEvidence checked