📊 Full opportunity report: ALIA. The Spanish answer. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Spain has announced ALIA, its largest publicly funded AI project, with a 40-billion-parameter multilingual model trained on 9.37 trillion tokens. Despite performance below Llama 2 benchmarks, it aims for widespread Spanish-speaking adoption and operational transparency.
Spain has officially launched ALIA, a 40-billion-parameter multilingual AI model trained on over 9 trillion tokens, funded entirely by public investment totaling €240 million. This project is a significant development in the broader context of hyperscaler investments and AI infrastructure. This marks the largest national AI project in Europe to date, with strategic emphasis on Spanish-language adoption and operational transparency. The project is led by the Barcelona Supercomputing Center and coordinated by Spain’s Secretary of State for Digitalisation and Artificial Intelligence, making it a key development in Europe’s sovereign AI landscape.
ALIA, which stands for “Artificial Linguistic Intelligence for Administration,” was developed with €90 million allocated for MareNostrum 5 upgrades and €150 million dedicated to integrating ALIA into Spanish industry. The model was trained from scratch on 12.875 trillion tokens, covering 35 European languages with a focus on Spanish and co-official languages, and was released under the Apache License 2.0 on HuggingFace on April 22, 2025. Despite its ambitious scope, benchmark results show ALIA’s performance lags behind Llama 2, with a 51.77% accuracy on XNLI in English versus Llama 2’s 66%, and 81.53% on SQuAD in English compared to Llama 2’s 93-94%. The project emphasizes multilingual coverage and transparency, validated by AESIA, but the performance gap underscores the importance of strategic investments in AI development.
Official statements from project leaders, including Josep M. Martorell, highlight that ALIA’s primary goal is widespread adoption within the Spanish-speaking world rather than achieving top benchmark scores. The project represents a strategic positioning aligned with Position 3 — focusing on operational relevance, regional adoption, and co-official language support, contrasting with the more performance-centric Position 1 models. The initiative is part of Spain’s broader national AI strategy, aiming to establish a sovereign AI infrastructure and foster local innovation, with a focus on operational transparency and multilingual capabilities.
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
Natural Language Processing with Transformers, Revised Edition
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.
Spanish AI chatbot
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.
Cursor AI for Programmers: Build, Debug, Refactor, and Ship Code Faster with AI Without Losing Control
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.
Rebooting the Machines: A New Human Vision for Artificial Intelligence
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications of ALIA for Europe’s AI Sovereignty
ALIA’s launch signifies Spain’s commitment to developing a sovereign AI infrastructure, emphasizing multilingual and regional adoption over raw performance. While benchmark results show a performance gap compared to models like Llama 2, the project’s focus on transparency, open-source licensing, and regional language coverage aligns with broader European strategies for AI independence. The substantial public investment underscores a strategic shift toward operational relevance and regional influence, positioning Spain as a key player in Europe’s AI landscape. However, the performance gap raises questions about the model’s immediate applicability for high-stakes tasks, emphasizing the importance of operational transparency over benchmarking supremacy.
Spain’s National AI Strategy and European Sovereign Models
Spain’s ALIA project is part of a broader European effort to develop sovereign AI models, following initiatives in Portugal, Italy, France, Germany, and Switzerland. These efforts aim to reduce dependence on US and Chinese models, foster regional innovation, and ensure AI transparency and control. Previously, projects like Portugal’s AMÁLIA, Italy’s Minerva, and France’s Mistral laid foundational work, often with smaller budgets and benchmarks. ALIA stands out as the largest publicly funded European national project, with €240 million dedicated to a 40B parameter model trained from scratch, emphasizing multilingual coverage and open-source licensing. The project’s emphasis on Spanish and co-official languages reflects Spain’s strategic focus on regional language support and operational transparency, aligning with European sovereignty goals.
“The goal is not to be the best-performing LLM in the world, but the most widely adopted in the Spanish-speaking world.”
— Josep M. Martorell
Performance and Operational Limitations Still Unclear
While benchmark results confirm ALIA’s performance below Llama 2, the practical implications for deployment in high-stakes or commercial contexts remain unclear. It is also uncertain how the model’s multilingual capabilities will perform across all 35 European languages in real-world applications, and whether the focus on transparency and regional adoption will translate into widespread operational use in Spain and beyond. Further testing and real-world deployment data are needed to assess its full impact.
Next Steps for ALIA’s Deployment and Evaluation
Future developments will include broader testing in governmental, industrial, and academic settings to evaluate ALIA’s operational readiness and multilingual performance, which ties into the ongoing discussions about AI market dynamics and investments. Additional benchmarks and case studies are expected to be released, along with updates on integration into Spanish industry and public services. The project team may also pursue further funding to enhance model capabilities or develop successor models aligned with operational and performance goals. Monitoring ALIA’s adoption and real-world impact over the coming months will be key to understanding its strategic success.
Key Questions
What is the main goal of Spain’s ALIA project?
The primary goal is to create a widely adopted, multilingual AI model focused on Spanish and regional languages, emphasizing operational transparency and regional influence over benchmark performance.
How does ALIA compare to other European AI models?
Benchmark results show ALIA performs below models like Llama 2 in standard tests, but it is the largest publicly funded European national project, with a focus on regional language support and transparency.
What are the main limitations of ALIA currently?
Benchmark performance lags behind leading models, and its real-world operational capabilities, especially in high-stakes contexts, remain to be fully tested and demonstrated.
Will ALIA be open for commercial use?
Yes, ALIA was released under the Apache License 2.0, making it open-source and accessible for further development and deployment, though its commercial viability depends on future performance and integration.
What does this mean for Europe’s AI sovereignty efforts?
ALIA signifies a major step in European efforts to develop sovereign, regionally focused AI models, emphasizing transparency and regional adoption over benchmark dominance, aligning with broader European strategic goals.
Source: ThorstenMeyerAI.com
