The first “large language model” of Artificial Intelligence in Spanish was made available to users in Spain this Monday, announced the Government led by Pedro Sánchez. “The first models of Alia, the foundational model of Artificial Intelligence in Spanish, were published a few minutes ago,” said Sánchez himself, at a morning conference in Madrid dedicated precisely to Artificial Intelligence (AI).

Sánchez emphasized that these are “public and open” AI models, which can begin to be used as of this Monday. Alia (known as “ChatGPT in Spanish”) is a “large language model” (LLM) for Castilian and other official languages ​​in Spain (Galician, Catalan, Valencian and Basque).

As it is an open source tool, it can be used, for example, by companies as a starting point to develop their own models. The Spanish Government itself will move forward, for now, with two pilot projects, which aim to apply Alia in the national Tax Agency and in the health sector.

In the first case, the objective is to streamline the agency's work and, in the second, to improve the diagnosis of heart failure, according to the Government. Sánchez's executive had already announced last year that Spain would have the first LLM trained in Spanish – and in the country's other official languages ​​–, the Alia, whose first models were launched this Monday.

At the conference in Madrid, Pedro Sánchez announced, in parallel, 150 million euros to promote the use of AI in Spanish companies, 20 of which are destined for small and medium-sized companies. Sánchez defended the potential and opportunities of AI, but highlighted the need for application and responsible use.

In May last year, the Spanish Government approved a new “Artificial Intelligence Strategy 2024-2025”, which foresees an investment of 1.5 billion euros. The Portuguese LLM, which will be called Amália, was announced in November last year by Prime Minister Luís Montenegro and is due to be launched in 2026.