📊 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 but raises critical questions about openness, native-language data sufficiency, and optimization goals. These issues influence broader European sovereign-LLM strategies.
Portugal’s €5.5 million AMÁLIA large language model (LLM) is now operational, with its base version released and benchmarks showing strong performance on European Portuguese tasks. However, critical questions about its openness, native-language data, and strategic goals remain unaddressed, raising concerns about the broader European sovereign-LLM movement.
The AMÁLIA project, involving around 60 researchers from Portugal’s top research institutions, was announced in December 2024 and completed its base version by September 2025. The model, designed to handle Portuguese text, outperforms previous open models on several benchmarks and is accessible to 450,000 academic users via the FCT’s IAedu platform. It is built as a continuation of the EuroLLM multilingual foundation, incorporating approximately 5.8 billion tokens from Portuguese web archives during extended pre-training, and about 17-18% of supervised fine-tuning data is in European Portuguese.
Despite these technical achievements, Duarte O.Carmo’s analysis highlights that fundamental questions remain unanswered: How open is ‘fully open’? How much native-language data is enough? And what should be the primary optimization goals? These questions are central to evaluating the model’s strategic and policy relevance, especially as similar efforts emerge across Europe, including Italy’s Minerva, Germany’s Aleph Alpha, and France’s Mistral.
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.
European Portuguese language large language model
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 research platform for Portuguese language
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.
open-source LLM development tools
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.
AI benchmark datasets for Portuguese
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-Language AI Strategies
The questions raised by the AMÁLIA project are critical for shaping national AI policies across Europe. The model exemplifies the challenges of balancing openness, native-language data, and strategic objectives in sovereign-LLM development. How these questions are answered will influence future investments, data policies, and the international standing of European AI efforts, especially amid increasing competition from commercial and non-European models.
European Sovereign-Language Model Initiatives and Challenges
Across Europe, several countries are investing in sovereign-language LLMs, including Italy’s Minerva, Germany’s Aleph Alpha, and France’s Mistral. These efforts aim to foster national AI independence and linguistic sovereignty. However, they face common structural challenges: defining openness, sourcing sufficient native-language data, and setting clear objectives for model deployment and strategic use. Portugal’s AMÁLIA is a prominent case illustrating these issues, with its public funding and national scope making the questions particularly urgent.
“The AMÁLIA project raises fundamental questions about openness, native data, and strategic goals that the European sovereign-LLM movement must address openly.”
— Duarte O.Carmo
Unanswered Questions About AMÁLIA’s Openness and Strategy
It is not yet clear how open the AMÁLIA model truly is, especially regarding access to training data and model weights. Additionally, the optimal amount of native-language data and the primary goals for the model’s deployment are still under discussion. The final version, expected in June 2026, may address some of these gaps, but current details remain uncertain.
Next Steps for Portugal’s AMÁLIA and European Sovereign LLMs
The final version of AMÁLIA is scheduled for release in June 2026, which may clarify some of the current uncertainties around data, openness, and objectives. Meanwhile, ongoing evaluations and policy debates across Europe will likely intensify, influencing future investments and strategic directions for national LLM initiatives.
Key Questions
What are the main technical features of AMÁLIA?
AMÁLIA is built as a continuation of the EuroLLM multilingual foundation, incorporating approximately 107 billion tokens during extended pre-training, with about 5.8 billion tokens from Portuguese web archives, and has shown strong performance on Portuguese benchmarks.
Why are questions about openness and native data important?
These questions determine how accessible the model is, influence data privacy and sovereignty, and shape strategic policy decisions about national AI independence and collaboration.
How does AMÁLIA compare to other European models?
Currently, AMÁLIA outperforms most open models on Portuguese benchmarks but still lags behind some proprietary models like Qwen 3-8B on certain tasks. Its strategic approach differs by building on a multilingual foundation rather than training from scratch.
What are the broader implications for European AI policy?
The case of AMÁLIA highlights the need for clear policies on model openness, native-language data sourcing, and strategic goals, which will influence the future of AI sovereignty across Europe.
When will the final version of AMÁLIA be available?
The final version is scheduled for release in June 2026, at which point more definitive answers to current uncertainties are expected.
Source: ThorstenMeyerAI.com
