📊 Full opportunity report: ALIA. The Spanish answer. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Spain’s ALIA project, funded with over €240 million, has released the Salamandra-40B model trained on 9.37 trillion tokens. While it aims to be Europe’s first public multilingual LLM, benchmark results show it underperforms compared to Llama 2, highlighting structural capability gaps.
Spain’s ALIA project has publicly released the Salamandra-40B language model, marking the country’s most ambitious effort yet in developing a sovereign, multilingual artificial intelligence system. Funded with over €240 million from public sources, ALIA aims to position Spain as a leader in European AI sovereignty, emphasizing Spanish and co-official languages. Despite its strategic goals, benchmark data indicates its performance remains below that of leading models like Llama 2.
The ALIA initiative, managed by the Barcelona Supercomputing Center (BSC-CNS) and led by the Secretary of State for Digitalisation and Artificial Intelligence (SEDIA), was launched with €90 million dedicated to upgrading the MareNostrum 5 supercomputer and an additional €150 million allocated for integrating ALIA into industry and government applications. The Salamandra-40B model, trained on 9.37 trillion tokens across 35 European languages and 92 programming languages, was released under the Apache License 2.0 on HuggingFace on April 22, 2025.
Benchmark results reveal that ALIA-40B scores approximately 51.77% on the XNLI_en test and 81.53% on SQuAD_en, both below Llama 2’s performance of around 66% and 93-94% respectively. These figures confirm a structural capability gap, especially at the 40B scale, indicating that ALIA’s current performance does not match the leading models. The project emphasizes Spanish language and multilingual coverage, with a strategic focus on widespread adoption within the Spanish-speaking world rather than outperforming global benchmarks.
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.
Mini AI Voice chatbot, smart Voice Assistant, Multiple AI Models, Emotional Interaction, 100+ Stickers, Suitable for Home and Office use, (Black)
1. Emotional Interaction: This chatbot can recognise and respond to your emotions, offering a more personalised and human-like…
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.
Beyond the Public Cloud: Architecting Private, Secure, and Sovereign AI for the European Enterprise
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.
Applying AI in Learning and Development: From Platforms to Performance
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications for European AI Sovereignty
Spain’s ALIA project exemplifies Europe’s broader effort to develop sovereign AI infrastructure, reducing reliance on US and Chinese models. This includes the ongoing discussion about hyperscaler CapEx and AI investments. Its emphasis on Spanish and co-official languages aims to foster regional adoption and linguistic diversity, aligning with EU policies promoting digital sovereignty. However, benchmark results suggest that while the project advances strategic positioning, it faces operational limitations in raw performance, raising questions about the balance between strategic goals and technological capabilities.
Spain’s Strategic Position in European AI Development
Prior to ALIA, European efforts included projects like Portugal’s AMÁLIA, Italy’s Minerva, and pan-European initiatives such as OpenEuroLLM and Mistral, with varying scales and funding sources. ALIA is the largest publicly funded national project in Europe, with over €240 million dedicated to training a 40B parameter multilingual model from scratch, operating within a broader context of EU-led sovereignty ambitions. The project is part of Spain’s broader national AI strategy, launched publicly in January 2025, aiming to build a sovereign AI ecosystem aligned with European policies.
Benchmarking and structural analyses indicate that ALIA’s operational performance is below that of international models like Llama 2, reflecting a capability gap at the current scale. The project leadership emphasizes widespread adoption over top-tier performance, framing ALIA as a strategic tool for Spanish and European digital independence.
“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, ALIA project lead
Operational Performance vs. Strategic Goals Unclear
Benchmark data confirms ALIA-40B’s performance lags behind leading models like Llama 2, but it remains unclear how this performance gap will evolve with further development or additional training. The project’s long-term impact on European AI sovereignty and industry adoption also remains to be seen, as operational benchmarks are only one aspect of strategic success.
Future Development and Adoption Trajectory
Next steps include ongoing model fine-tuning, expanding multilingual capabilities, and increasing adoption within Spanish and European industries. The project team also plans to release additional models and benchmarks, aiming to improve performance and demonstrate operational viability. Monitoring how ALIA’s strategic positioning influences European AI policy and industry uptake will be key in the coming months.
Key Questions
What is the main goal of Spain’s ALIA project?
The primary goal is to develop a widely adopted, sovereign multilingual AI model tailored to Spanish and European languages, emphasizing regional adoption over outperforming global models.
How does ALIA compare to other models like Llama 2?
Benchmark results show ALIA-40B scores below Llama 2 in key tests, indicating a structural capability gap at this scale, though ALIA aims for strategic regional relevance rather than top-tier performance.
What are the funding sources for ALIA?
ALIA is funded entirely through public investment, totaling over €240 million, including €90 million for supercomputer upgrades and €150 million for model development and industry integration.
What does the performance gap imply for ALIA’s future?
The performance gap suggests that while ALIA advances Spain’s strategic sovereignty goals, it may face limitations in operational capabilities compared to leading international models, affecting its competitiveness and adoption.
Source: ThorstenMeyerAI.com
