Name in original language
Latam‑GPT
Initiative overview
LatamGPT is a collaborative initiative designed to strengthen artificial intelligence capabilities across Latin America and the Caribbean. Although its most visible output is the development of an open large language model (LLM), the project encompasses a much broader effort to build the foundations of a regional AI ecosystem. It combines model development with investments in data, infrastructure, evaluation tools, research collaboration and human capital.
A central objective of LatamGPT is to improve the representation of the region in AI systems. Many leading AI models are trained primarily on data originating from the Global North and may not adequately reflect the linguistic, cultural and social realities of Latin America and the Caribbean. LatamGPT seeks to address this gap by incorporating data, knowledge and perspectives that are more representative of the region's diverse countries, languages and cultures. The goal is to enable AI systems that perform more effectively on topics and contexts that matter to local users.
The initiative also aims to strengthen regional technological capabilities. Rather than focusing solely on the use of AI tools developed elsewhere, LatamGPT seeks to foster expertise in the development, training and evaluation of advanced AI systems. Through collaboration among researchers, universities, technology institutions and public organisations, the project helps build technical skills and supports the emergence of a stronger regional innovation ecosystem.
Open access is another key component of the project. The model and many supporting resources are intended to be openly available, allowing researchers, companies, governments and civil society organisations to adapt and use them for their own purposes. This approach is intended to lower barriers to access, encourage innovation and enable a wider range of actors to develop AI applications tailored to regional needs.
In addition to the model itself, LatamGPT includes complementary initiatives that support the broader AI ecosystem. These include the creation of regional datasets, the development of benchmarks to evaluate AI systems on Latin American knowledge and culture, and the collection of user interactions to improve future model versions. Together, these activities contribute to a shared pool of knowledge and resources that can benefit stakeholders across the region.
Overall, LatamGPT is best understood not only as an AI model, but as a strategic regional effort to build the capabilities, infrastructure and expertise needed for long-term AI development. By combining open technology, local knowledge and regional collaboration, it aims to support innovation, competitiveness and greater participation of Latin America and the Caribbean in the development of advanced AI systems.
