Lara Mini Base: Language Diversity as a Resource, Not a Trade-off
cs.CL, cs.AI
Submitted: 2026-11-12
Updated: 2026-11-12
Comments: 51 pages, 14 figures, 29 tables. Model weights: https://huggingface.co/translated/Lara-4-Mini-Base
Code: https://github.com/deep-spin/EuroMoE
License: http://creativecommons.org/licenses/by-nc-sa/4.0/
The gist: General-purpose language models are trained mostly on English: over open models we review, at most, non-English pretraining data makes up only a 12% share.
Abstract
General-purpose language models are trained mostly on English: over open models we review, at most, non-English pretraining data makes up only a 12% share. Models built for European languages give them far more, but score well below general-purpose models of the same size. A common concern is that, in small models, languages compete for capacity. We present Lara Mini base, a model trained from scratch on 6T tokens in English and 44 other languages, which together cover all 24 official languages of the European Union. As training progresses, the 44 other languages together come to outweigh English. To give these languages enough capacity at low cost, Lara Mini base uses a mixture of experts architecture with 7.2B total and 0.6B active parameters. Among open base models with up to 2B active parameters, it achieves the highest average score across standard English, multilingual, math, and code benchmarks, ahead of models trained on 4 to 6 times as many tokens. Its margin is largest on multilingual benchmarks in the languages it supports and on math benchmarks. It supports contexts of up to 128k tokens. We find that our MoE architecture reduces capacity competition between languages. For a single user on one GPU, without quantization, it generates text at least 1.8 times faster than any model with a higher average score. We release the weights of Lara Mini base under the CC BY-NC-SA 4.0 license. In addition, we will release T-HellaSwag, a higher-quality machine translation of the commonsense reasoning benchmark HellaSwag into 110 language variants.
Sources
- Kolibri: A Sovereign European Model on the Pareto Frontier
- LFM2 Technical Report
- Program Synthesis with Large Language Models
- Longformer: The Long-Document Transformer
- Lessons from the Trenches on Reproducible Evaluation of Language Models
- Does your data spark joy? Performance gains from domain upsampling at the end of training
- Global PIQA: Evaluating Commonsense Reasoning Across 100+ Languages and Cultures
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- Training Verifiers to Solve Math Word Problems
- Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models
- DeepSeek-V3 Technical Report
- Falcon-H1: A Family of Hybrid-Head Language Models Redefining Efficiency and Performance
- Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining
- Gemma: Open Models Based on Gemini Research and Technology
- Gemma 3 Technical Report
- Gemma 4 Technical Report
- Salamandra Technical Report
- The Llama 3 Herd of Models
- Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data
- Mistral 7B
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data