The Portuguese large-scale language model (LLM) Amália, designed as a counterpart to ChatGPT, will not be available to the general public until the end of the first half of 2026. Currently, the beta version is only accessible to the research centers involved in its development.
The completion of the beta phase of Amália, initially scheduled for the end of the first quarter of 2025, will not be made available to the public. The model’s development is being carried out by a consortium that includes the Universidade Nova de Lisboa's Nova LINCS research centers, Instituto de Telecomunicações, Instituto Superior Técnico, and other national institutions known for their contributions to AI research. The goal is to refine the model with the collaboration of various research centers, focusing on specific domains such as media, science, and education.
Funded with €5.5 million from Portugal’s Recovery and Resilience Plan (RRP), the project has a phased development plan spread over 18 months. The final version of Amália, a multimodal model capable of interpreting various data formats like text, images, and videos, is expected to be ready by the second quarter of 2026.
Once completed, Amália will be released in an open-source format, free to the public. The model will be available for use by civil society—including companies, research centers, and the public administration—enabling the development of applications like educational tutors or virtual assistants for government services. The model’s ultimate goal is to strengthen Portugal's technological capabilities, with the public sector playing a key role in its deployment.
In addition to advancing AI research, the Amália project is part of Portugal's broader National Agenda for Artificial Intelligence, though its formal presentation has been delayed due to recent political changes.
Related News