📊 Full opportunity report: AMÁLIA · The Three Hard Questions. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Portugal’s AMÁLIA, a €5.5 million European Portuguese LLM, is operational and shows strong performance, but experts raise three fundamental questions about its openness, native data, and objectives. The final version is due June 2026.
Portugal’s €5.5 million AMÁLIA large language model is now operational, demonstrating superior performance on Portuguese benchmarks compared to previous models, but fundamental questions about its openness, native-language data, and strategic goals remain unanswered.
The project involves around 60 researchers across Portugal’s leading academic institutions, including NOVA, IST, and IT, and was announced in December 2024. The base model was completed by September 30, 2025, and is currently accessible via the FCT’s IAedu platform to 450,000 academic users. It handles text only, with multimodal features planned for future versions. The model is a continuation of the EuroLLM multilingual foundation, not trained from scratch, contrasting with Italy’s Minerva, which was trained from the ground up on Italian and English data.
Performance benchmarks show AMÁLIA surpasses previous open models on European Portuguese tasks and beats Qwen 3-8B on most benchmarks, though it still trails Qwen on some specific tests like ALBA. The model’s training involved 107 billion tokens, with a small proportion (about 5.8 billion) from Portuguese web archives, and approximately 17-18% of supervised fine-tuning data was Portuguese. The final version is scheduled for release in June 2026, with ongoing evaluations and potential improvements.
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.
Portuguese Flash Cards – Learn Portuguese Language Vocabulary Words and Phrases – Basic Language for Beginners – Gift for Travelers, Kids, and Adults by Travelflips
- Language: Portuguese flash cards for beginners
- Learning Aid: Phonetic pronunciation and English translation
- Quality: Durable cards with portable box
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.
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.
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.
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-Language AI Strategies
The development of AMÁLIA exemplifies a broader European effort to create national AI models, highlighting the importance of transparency, native-language data, and strategic objectives. The questions raised about openness and data reflect core policy concerns about sovereignty, data privacy, and technological independence, making this case a critical reference point for policymakers and researchers across Europe.
While the model demonstrates technical success, the unresolved questions about openness, native data sufficiency, and goal-setting could influence future funding, regulation, and strategic directions for European AI initiatives. The way these questions are addressed will shape Europe’s position in the global AI landscape and inform the development of models that truly reflect national languages and cultures.
European Sovereign-Language LLM Efforts and Challenges
Across Europe, multiple countries are investing in national large language models, including Italy’s Minerva, Germany’s Aleph Alpha, France’s Mistral, and others. These projects share a common challenge: balancing openness with strategic control, ensuring sufficient native-language data, and defining clear objectives for their models. The European sovereign-LLM movement is still in its early stages, often operating without explicit answers to these fundamental questions, which impacts transparency and strategic coherence.
Portugal’s AMÁLIA is a key case study because of its public funding and national scope, illustrating how these issues are not only technical but also political and strategic. The ongoing development of AMÁLIA reflects broader trends and challenges faced by the European AI community in establishing independent, culturally aligned language models.
“The three questions about openness, native data, and objectives are essential for evaluating the true progress and strategic direction of national LLMs.”
— Duarte O.Carmo
Unresolved Questions About Openness, Data, and Goals
While AMÁLIA is operational and performing well in benchmarks, it remains unclear how open the model truly is, given the limited transparency about its training data and licensing. The sufficiency of native Portuguese data for future improvements is also uncertain, as the current dataset is relatively small compared to the total training tokens. Additionally, the strategic objectives—whether the model aims for broad accessibility, commercial deployment, or national sovereignty—are not yet explicitly defined or publicly clarified.
These uncertainties are central to evaluating the model’s long-term impact and the broader European sovereign-LLM movement’s coherence, but definitive answers are still forthcoming.
Upcoming Milestones and Strategic Discussions
The final version of AMÁLIA is scheduled for release in June 2026, which will provide an opportunity to assess its capabilities, openness, and strategic alignment more comprehensively. Over the next 12-24 months, researchers and policymakers will likely scrutinize the model’s training data transparency, licensing, and deployment plans. Public discussions and potential regulatory frameworks are expected to evolve around these issues, shaping the future of European sovereign-language AI projects.
Additionally, ongoing benchmarking and community feedback will influence further development, with possible updates to the model’s architecture, data sources, and strategic goals based on emerging insights and policy debates.
Key Questions
What are the main concerns about AMÁLIA’s openness?
Experts question whether the model’s training data is sufficiently transparent and accessible, which is critical for assessing its independence and alignment with European data sovereignty principles.
How much native Portuguese data was used in training AMÁLIA?
Approximately 5.8 billion tokens from Portuguese sources, mainly web archives, were used in the extended pre-training, representing about 5.5% of the total training tokens.
What are the strategic goals for AMÁLIA?
The explicit strategic objectives remain unclear, but the model’s development appears aimed at national sovereignty, academic use, and benchmarking, with future plans for multimodal capabilities.
Why are these questions important for European AI development?
Addressing these questions ensures transparency, strategic coherence, and sovereignty, which are vital for Europe’s independent AI ecosystem and its global competitiveness.
Source: ThorstenMeyerAI.com
