Ctrl K
Contact
Use · EuroLLM

Is it good enough?

Ownership levelSubstantialnone·limited·partial·substantial·fullAnalytical input C ยท 69.2/100

This page is a projection of the one entry record, the Doesn't fail you factor that Use covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.

Which domain expands which factor
  • AssessUse & modify + Transparency
  • ImplementData control + Doesn't fail you
  • UseDoesn't fail you
  • SupportTransparency

Capabilities & modalities

EuroLLM is a text-only multilingual language model for European-language generation, assistant tasks and machine translation. Per aggregated benchmarks and the technical report (ev-tech-report, ev-moonlight) it is the strongest for its size on multilingual EU-language tasks - on par with Gemma-2-9B on multilingual benchmarks, ahead on WMT24++ translation, and matching Mistral-7B on English - while remaining, at 9B, below frontier on raw single-language reasoning. Choose EuroLLM when you need broad, even EU-language coverage from a European, Apache-2.0 model more than you need class-topping English reasoning.

Context window & long-context behaviour

Both EuroLLM-9B and EuroLLM-1.7B use a 4,096-token context window. This is short by current standards, so design around it:

  • Chunk long inputs and use retrieval rather than stuffing large documents into the prompt.
  • Budget the 4k across system prompt, few-shot examples, input and output.

OneHill has not independently measured effective long-context recall this session; treat 4k as the architectural ceiling.

Prompt format & chat template

The Instruct variants ship a chat template in their tokenizer configuration (with BOS token id 1 and EOS token id 4). Use it rather than hand-rolling role markers:

messages = [
    {"role": "system", "content": "You are a helpful multilingual assistant."},
    {"role": "user", "content": "Traduis en portugais : ..."},
]
inputs = tok.apply_chat_template(messages, tokenize=True,
                                 add_generation_prompt=True, return_tensors="pt")

Base checkpoints have no chat template - they are plain text-completion models and should not be prompted as chat assistants.

Language coverage

This is EuroLLM's defining strength. It is trained from scratch for multilingual EU coverage: all 24 official EU languages - Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Irish, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Slovak, Slovenian, Spanish, Swedish - plus roughly 11 additional strategic languages: Arabic, Catalan, Chinese, Galician, Hindi, Japanese, Korean, Norwegian, Russian, Turkish and Ukrainian.

Unlike English-centric open models where non-English behaviour is incidental, EU-language coverage is EuroLLM's design target, and translation quality (WMT24++) is a documented strength. The practical caveat: with 35 languages sharing a 9B (or 1.7B) budget, quality is uneven - the smaller and lower-resource languages trail the majors - so evaluate each target language for your task rather than assuming uniform quality.

How this scores

The ownership factor this domain covers, drawn from the one entry record.

3

Doesn't fail youIs it reliable and good enough for the job?

Moderate

Best-in-class multilingual for its size (on par with Gemma-2-9B, ahead on WMT24++ translation), but a 9B model with a short 4k context and light, uneven per-language safety coverage - a compact backbone, not a frontier or long-context engine.

How this scores (AOI sub-dimensions)
Performance3/5how capable it is relative to its classBest-in-class for its niche: independent and publisher benchmarks put EuroLLM-9B on par with Gemma-2-9B on multilingual EU-language tasks, ahead on WMT24++ translation, and matching Mistral-7B on English.
Operational4/5how practical it is to run, serve and maintain in productionStandard safetensors with a conventional architecture load on all mainstream stacks (vLLM, llama.cpp, Ollama, TGI, transformers); community GGUF / MLX / GPTQ quants and a hosted NVIDIA NIM endpoint exist.
Safety3/5whether misuse risks are evaluated and guardrails are providedThe Instruct variants are instruction-tuned on EuroBlocks and behave as ordinary safety-tuned assistants, but published safety/alignment coverage is light, there is no companion guard model, and multilingual coverage means safety behaviour varies by language and is not uniformly evaluated.
What this means for adoptionYou fully own the use of these Apache-2.0 weights - clean across 9B, 1.7B and 22B-Instruct-2512, with unconditional commercial use and no field-of-use limits - and, self-hosted and EU-domiciled, you keep your data entirely. It stays short of full because data openness is partial, not full: the SFT set (EuroBlocks-SFT-2512) is released but the ~4T-token pre-training corpus is only described, so you can inspect the recipe but not reproduce it from a released corpus as with a fully-open model; the 9B size, 4k context and uneven multilingual safety also cap reliability. Adopt it for multilingual EU-facing work that fits inside 4k, behind your own guardrails; look elsewhere for a long-context or frontier engine.

Sources

The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.

Vendor announcementunverified2026-07-25
EuroLLM is presented as an open-source European multilingual LLM suite covering all 24 official EU languages plus additional strategic languages, developed by an EU-funded consortium.
Model cardread2026-07-25
EuroLLM-9B-Instruct model card, read verbatim: hosted on the verified utter-project org on Hugging Face in safetensors under Apache-2.0, trained on 4T tokens, instruction-tuned on EuroBlocks, a 4k context and a documented chat template (BOS id 1 / EOS id 4); noted as not preference-aligned (may hallucinate or produce harmful content); comparable to Gemma-2-9B on EU languages and Mistral-7B on English.
Model cardread2026-07-25
EuroLLM-1.7B-Instruct model card, read verbatim: 1.7B parameters, a 4,096-token context, coverage of 35 languages, Apache-2.0, and comparable to Gemma-7B on machine-translation benchmarks.
Technical_reportread2026-07-25
EuroLLM-9B: Technical Report (arXiv 2506.04079), read verbatim: EuroLLM is trained from scratch to cover all 24 official EU languages plus 11 more, and documents the tokenizer, architecture, ~4T-token data mixture, EuroFilter pipeline, EuroBlocks-Synthetic post-training set and Megatron-LM training; the report frames its "Open release" as the "Public availability of models, filters, and datasets."
Documentationread2026-07-25
EuroLLM-22B-Instruct-2512 model card, read verbatim: openness is PARTIAL, not fully-open - the SFT dataset EuroBlocks-SFT-2512 IS released (utter-project/EuroBlocks-SFT-2512), but the ~4T-token pretraining corpus (web, parallel, Wikipedia, Arxiv, books, math, code, Apollo) is only described, not published as one downloadable dataset.
Vendor announcementunverified2026-07-25
The EuroLLM team's release blog describes EuroLLM-9B, its EU consortium and EU funding (EuroHPC / Horizon Europe), multilingual coverage, Apache-2.0 licensing and the release of the models, filter and synthetic post-training data.
Licenceread2026-07-25
EuroLLM base and Instruct weights - 9B, 1.7B and 22B-Instruct-2512 - are distributed under OSI-approved Apache-2.0 with unconditional commercial use and no field-of-use limits (the licence is not a preference-alignment/safety grant, and the cards warn the models may hallucinate).
Third-party analysisunverified2026-07-25
The European Commission's Open Source Observatory profiles EuroLLM as EU-funded pioneering European open-source AI released under an open licence.
Third-party analysisunverified2026-07-25
An independent review of the EuroLLM-9B technical report reports it on par with Gemma-2-9B on multilingual benchmarks, ahead on WMT24++ translation, and matching Mistral-7B on English.
Third-party analysisunverified2026-07-25
Independent coverage ranks EuroLLM among the leading European-developed open models for multilingual EU-language performance.
Third-party analysisunverified2026-07-25
EuroLLM safetensors checkpoints load on mainstream serving stacks (vLLM, llama.cpp, Ollama, TGI, transformers), with community GGUF/MLX/GPTQ quants and a hosted NVIDIA NIM endpoint available.