📊 Full opportunity report: AMÁLIA · The Three Hard Questions. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Portugal’s €5.5M AMÁLIA LLM, publicly launched in late 2025, outperforms many benchmarks but raises three key questions about openness, native data, and objectives. These issues impact Europe’s sovereign AI strategies.
Portugal’s €5.5 million investment in the AMÁLIA large language model has resulted in a publicly available, high-performing model, raising urgent questions about the openness, native data sufficiency, and strategic objectives of European sovereign AI projects.
AMÁLIA, developed by a consortium of approximately 60 researchers from Portugal’s top institutions, was completed in September 2025 and made accessible through the FCT’s IAedu platform. It is based on a continuation of the EuroLLM multilingual foundation, with a focus on Portuguese, and currently handles text-only tasks, with multimodal capabilities planned for future versions.
Benchmarks show AMÁLIA surpasses previous open models on European Portuguese tasks and outperforms Qwen 3-8B on most benchmarks, though it still trails Qwen on some specific tasks like ALBA, the primary Portuguese benchmark. The project’s final version is scheduled for release in June 2026, with ongoing assessments of its capabilities and limitations.
AMÁLIA
The three hard
questions.
Portugal spent €5.5M to build a European Portuguese LLM. The base version is operational, the benchmarks beat Qwen 3-8B on most pt-PT tasks. So why are the most important questions still unanswered?
Last month, Duarte O.Carmo published the sharpest public analysis of AMÁLIA — Portugal’s state-funded European Portuguese large language model. He prefaces his critique with the necessary diplomatic apparatus before doing what almost nobody else in the European-sovereign-LLM discourse has been willing to do publicly: asking hard questions about whether the work, as released, actually does what it set out to do. This piece is a structural extension of his analysis. The AMÁLIA case study exposes three hard questions every national LLM effort needs to answer publicly — and the broader European sovereign-LLM movement has been operating without explicit answers to any of them.
Three questions every national LLM effort needs to answer publicly.
Duarte O.Carmo’s framing maps cleanly onto the structural argument. Each question lands specifically in AMÁLIA — and the broader European sovereign-LLM movement has been operating without explicit answers to any of them.
The three questions form a structural feedback loop. Q3 (optimization target) determines Q2 (data volume needed) which conditions Q1 (openness sufficient for community contribution). The European sovereign-LLM movement collectively benefits from these questions becoming standard methodology disclosure, not exceptional critique.
Advanced Language Tool Kit: Teaching the Structure of the English Language
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
107 billion tokens. 5.8 billion clearly pt-PT.
The structurally tractable question with a structurally surprising answer. For a model whose entire stated purpose is European Portuguese prioritization, the native-language share of extended pre-training is 5.5%. The implications cascade into every other question.
AI Voice Recorder, Note Pro Voice Recorder Transcribe & Summarize, AI Noise Cancellation Technology, Supports 152 Languages, 64GB Memory APP Control Audio Recorder for Lectures, Meetings, Calls
Cutting-Edge AI Transcription & Summarization: Leverage GPT-4o’s advanced intelligence in this top-tier AI voice recorder for real-time, highly…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Olmo standard. AMÁLIA’s current state.
Allen Institute for AI’s Olmo project defines what “fully open” operationally requires. Olmo doesn’t lead frontier benchmarks. That’s not the point. The point is to be the structural reference for openness. AMÁLIA’s “fully open source” claim should track to the operational standard.
Brazilian Portuguese Vocabulary 75 – Guided Language Learning Set with Audio | Ages 13+ | Essential Words & Sentences for Structured Study, Travel & Classroom Use
FOR AGES 13 AND UP – TEEN & ADULT LANGUAGE LEARNERS Designed for learners ages 13+, this guided…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Four strategic positions. AMÁLIA between two and three.
Approximately €100M+ in publicly disclosed European sovereign-LLM funding across the major initiatives. The structural question every project faces: what is the actual competitive position you’re staking? Four options — none mutually exclusive — but each requiring different commitments.
OpenAI Evals Cookbook: Designing Benchmarks for Product‑Grade LLM Features
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Three standards. For AMÁLIA and the movement.
The structural critique generalizes beyond AMÁLIA. Italy, France, Germany, Switzerland, the OpenEuroLLM consortium, and every subsequent national project benefit from public discourse holding national LLM efforts to operational standards on openness, data accounting, and strategic positioning.
The European sovereign-AI agenda is a serious strategic project that deserves serious public discourse. O.Carmo’s analysis is what serious public discourse looks like. Appropriately diplomatic. Structurally rigorous. Willing to ask the hard questions in public when the public investment justifies it. More of this is needed — across every European sovereign-LLM project, not just AMÁLIA.
Implications for European Sovereign AI Strategies
The development and public release of AMÁLIA highlight critical structural issues facing Europe’s AI ambitions, especially regarding transparency, native-language data adequacy, and strategic priorities. These questions are vital for shaping future investments and policies in sovereign language models, influencing how Europe competes in the global AI landscape.
European Sovereign LLM Initiatives and Structural Challenges
Across Europe, multiple countries and consortia—including Italy’s Minerva, Germany’s Aleph Alpha, France’s Mistral, and the OpenEuroLLM initiative—are pursuing large language models with varying approaches. A common challenge is balancing openness, native data utilization, and defining clear objectives. Portugal’s AMÁLIA exemplifies this broader trend, with its public funding and national scope making its structural questions especially pertinent.
While these projects aim to foster sovereignty and innovation, they often operate without explicit answers to core questions about how open models should be, what native data sufficiency looks like, and what strategic goals they should prioritize. Experts like Duarte O.Carmo have emphasized that these issues are underexamined but crucial for the future of European AI.
“The three questions—how open is ‘fully open,’ how much native data is enough, and what should we be optimizing for—are fundamental to understanding the true potential and limitations of Europe’s sovereign LLM efforts.”
— Duarte O.Carmo
Unanswered Questions About AMÁLIA’s Openness and Goals
It remains unclear how open AMÁLIA truly is beyond the publicly released version, especially regarding proprietary components or future multimodal capabilities. Additionally, the strategic objectives guiding its development—such as native data sufficiency and long-term goals—are still under discussion and have not been formally clarified by the team or policymakers.
Upcoming Milestones and Evaluation of AMÁLIA
The final version of AMÁLIA is scheduled for release in June 2026, which will provide a more comprehensive assessment of its capabilities and limitations. Over the next 12-24 months, researchers and policymakers will scrutinize its performance, openness, and strategic alignment, shaping Europe’s approach to sovereign language models.
Key Questions
What are the main concerns about AMÁLIA’s openness?
It is not yet clear how transparent the model’s underlying data and architecture are, especially regarding proprietary or restricted components, which affects trust and replicability.
How much native Portuguese data was used in AMÁLIA’s training?
Approximately 5.8 billion tokens from Portuguese sources, representing about 5.5% of the extended pre-training mixture, with supervised fine-tuning involving roughly 17-18% Portuguese data.
Why do these questions matter for Europe’s AI future?
Addressing these issues is crucial for ensuring transparency, strategic clarity, and competitive sovereignty in AI development across Europe.
What are the risks of not answering these structural questions?
Without clear answers, European models may face issues of trust, limited scalability, and strategic misalignment, potentially weakening their global competitiveness.
Source: ThorstenMeyerAI.com
