📊 Full opportunity report: ALIA. The Spanish answer. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Spain has launched ALIA-40B, a public-funded multilingual language model trained on 9.37 trillion tokens. While it emphasizes Spanish and co-official languages, benchmark results show it underperforms compared to Llama 2, highlighting a structural capability gap. The project exemplifies Spain’s strategic focus on widespread adoption over top-tier performance.
Spain has publicly launched ALIA-40B, a multilingual language model trained on over 9.37 trillion tokens, marking the country’s largest public AI initiative and its answer to the European sovereign-AI challenge.
Developed under the Spanish government’s institutional AI program, ALIA-40B was trained on the MareNostrum 5 supercomputer, utilizing 4,480 NVIDIA H100 GPUs. It covers 35 European languages with an oversampling of Spanish, aiming to promote Spanish-language adoption and co-official language support.
Funded entirely by public investment totaling over €240 million, the project is led by the Barcelona Supercomputing Center (BSC-CNS) and coordinated by the Secretary of State for Digitalisation and Artificial Intelligence (SEDIA). The model was released under the Apache License 2.0 on HuggingFace on April 22, 2025, and has undergone validation by AESIA.
Benchmark results indicate the model’s performance is below that of Llama 2, with 51.77% accuracy on XNLI in English compared to Llama 2’s 66%, and 81.53% on SQuAD in English versus Llama 2’s 93-94%. These results confirm a structural capability gap, aligning with prior analyses suggesting Position 3 strategic positioning—focused on multilingual, Spanish-centric adoption—is operationally more credible than the Position 1 performance claim.
ALIA.
The Spanish
answer.
€240M+ Spanish public funding · ALIA-40B + Salamandra family · 9.37T tokens · 35 European languages + 92 programming languages · MareNostrum 5 · Apache 2.0 release. The largest publicly funded European national-AI project by cumulative scope — and the empirical test case for the Position 1 vs Position 3 strategic-positioning argument.
This is the tenth standalone essay in the European sovereign-LLM track and the third Tier 2 expansion piece. ALIA is Spain’s institutional answer — the largest EU member state by GDP not yet documented in the track. The project markets itself as Position 1 + Position 2 simultaneously — “Europe’s first public multilingual foundational model.” The benchmark evidence (ALIA-40B 51.77% XNLI_en vs Llama 2 66%) confirms the structural capability gap from Finding 1 of the synthesis essay. The Position 3 framing — Martorell’s “most widely adopted in the Spanish-speaking world” — is operationally honest. €90M MareNostrum 5 upgrade + €150M company integration = €240M+ cumulative scope. Apache 2.0 open-source release + AESIA validation + co-official languages oversampling. Both can be true at once. The Spanish public discourse would benefit from explicit Position 3 strategic positioning.
Six models. Apache 2.0.
The ALIA family operates as a tiered model portfolio. ALIA-40B is the flagship at 40 billion parameters; the Salamandra family scales down to 7B, 2B and instruct-tuned variants; mRoBERTa provides the foundational multilingual baseline. All released under Apache License 2.0 on April 22, 2025 at the HispanIA 2040 event — “Public Code, Public Money” approach.
multilingual
MN5 LLM
edge
target
instruct
encoder
Multilingual AI Translation Mastery: Building Accurate, Culturally Sensitive Language Tools and Global Communication Systems in 2026
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Four official. Oversampled by factor of 2.
ALIA’s distinctive multilingual coverage strategy. The four co-official Spanish languages are oversampled by factor of 2 in the training corpus — structurally distinct from Apertus’s broad 1,811-language coverage approach. The strategy targets deep coverage of Spanish co-official languages rather than maximum language breadth.
FancyDove AI Assistant Device Powered by ChatGPT, No Subscription Needed, Standalone AI Chatbot Translator, AI Tutor for Learning, Writing & Homework, Portable AI Gadget for Students & Travel Black
No Subscription & Lifetime Access – Pay Once, Use AI Forever: Enjoy powerful AI chat, writing, translation, and…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
ALIA-40B vs Llama 2. 14-point gap.
The empirical evidence Finding 1 of the synthesis essay needed. ALIA-40B at 40 billion parameters with €240M+ public funding and 8+ months MareNostrum 5 training achieves performance below Llama 2 — a 2023 frontier model released approximately 18 months before ALIA-40B. The capability gap is real and consistent with six of seven prior national-project answers documented in the track.
The Cranky Man's Guide to LoRA & QLoRA: Personal Lessons from a Thousand LLM Fine-Tuning Fails
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Two pilots. Public administration deployment.
The operational deployment targets that validate the Position 3 + Position 4 framing. Public administration deployment is the structurally credible Position 3 + Position 4 strategic positioning — captive demand from Spanish public institutions where Spanish-language specialization is operationally distinctive.
The work is real across the Spanish ALIA case. €240M+ public funding committed. 40B parameter from-scratch model trained on 9.37 trillion tokens. Salamandra family released under Apache 2.0. AESIA validation aligned with EU AI Act transparency standards. Two pilot applications shipped — Tax Agency chatbot and primary care medicine heart failure diagnosis. The Position 1 framing is operationally misleading. ALIA-40B performance below Llama 2 confirms the structural capability gap. The Position 3 framing is operationally honest — Spanish-speaking world adoption, co-official languages oversampling, public administration deployment. Both can be true at once. The Spanish public discourse would benefit from explicit Position 3 strategic positioning.
public-funded AI tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications of ALIA’s Strategic Positioning
While ALIA-40B demonstrates Spain’s commitment to developing a large-scale, multilingual AI model with open-source accessibility and validation, its benchmark performance reveals a clear capability gap compared to leading models like Llama 2. This underscores Spain’s strategic focus on widespread adoption within the Spanish-speaking world and co-official languages, rather than competing for top performance globally.
The project exemplifies a broader European approach to national AI sovereignty, emphasizing language coverage, transparency, and public ownership over raw performance metrics. It also highlights the ongoing tension between strategic positioning—Position 1 (performance-driven) versus Position 3 (adoption and language focus)—with ALIA aligning more with the latter.
Background on Spain’s AI Strategy and ALIA Development
Spain’s ALIA project is part of a broader national effort to establish sovereign AI capabilities, supported by €240 million in public funds and coordinated by the Barcelona Supercomputing Center. It follows a series of European national models, including Portugal’s AMÁLIA, Italy’s Minerva, and the pan-European OpenEuroLLM, but is distinguished as the largest publicly funded initiative in Europe by scope and scale.
The project originated from Spain’s 2019 Language Technologies Plan and was further advanced with the 2024 Spanish AI Strategy, emphasizing multilingual coverage and public ownership. The training utilized MareNostrum 5’s high-performance computing resources, with the goal of fostering Spanish-language AI adoption and supporting co-official languages across Europe.
Previous assessments suggested that models trained with a focus on multilingual coverage, especially Spanish, often lag behind performance benchmarks like Llama 2, which is performance-optimized but less language-diverse. ALIA’s benchmark results reinforce this pattern, confirming the structural challenge of balancing language coverage and performance.
“Our goal is not to be the best-performing LLM in the world, but to create the most widely adopted model in the Spanish-speaking world.”
— Josep M. Martorell, ALIA project lead
Unresolved Questions About ALIA’s Performance and Impact
It remains unclear how ALIA-40B will perform in real-world applications beyond benchmark tests, especially in Spanish and co-official languages. The long-term adoption rate and integration into industry and government sectors are still to be observed.
Additionally, the strategic implications of the performance gap versus the language and transparency focus are ongoing debates within European AI policy discussions.
Next Steps for ALIA and Spain’s AI Ambitions
Further benchmarking and real-world testing will clarify ALIA-40B’s practical capabilities. The project team plans to expand multilingual support and industry integration, aiming for wider adoption within Spain and potentially across Europe.
Monitoring how the model influences Spain’s position in European AI sovereignty debates and its role in fostering local AI ecosystems will be critical in the coming months.
Key Questions
What is the main goal of Spain’s ALIA project?
The main goal is to develop a multilingual, publicly funded AI model that promotes Spanish-language adoption and supports co-official languages, focusing on widespread use rather than top-tier performance.
How does ALIA-40B compare to other models like Llama 2?
Benchmark results show ALIA-40B performs below Llama 2 in key tests, indicating a structural capability gap. Its focus is on multilingual coverage and transparency rather than raw performance.
What are the strategic implications of ALIA’s development?
It exemplifies Spain’s positioning within European AI policy as prioritizing language coverage, openness, and adoption over competing for the highest performance benchmarks.
Will ALIA-40B be used commercially or mainly for government purposes?
The project aims for broad adoption within Spain and the Spanish-speaking world, including integration into government, industry, and public services, but commercial deployment details are still emerging.
Source: ThorstenMeyerAI.com
