Back
ArchaeologyInnovations

OeAW: Apollo AI is now learning Latin

OeAW researcher Dolganov wins highly competitive grant from US organization Schmidt Sciences to establish a new Large Language Model for a historical language

15.10.2026
Latin tablets in the Grand Egyptian Museum, Cairo.
© Anna Dolganov

Millions of historical documents, manuscripts, and printed editions in Latin, from antiquity to the modern era, are held in collections around the world. Given the sheer volume of these sources, many have yet to be systematically explored or made accessible to researchers. By developing the first Large Language Model (LLM) specifically designed for Latin, the Austrian Academy of Sciences (OeAW) aims to unlock access to these sources and enable new discoveries across nearly three millennia of history.

The project is now attracting significant international attention: Anna Dolganov, researcher at the Austrian Archaeological Institute of the OeAW, has secured a grant of 810,000 USD from the Humanities and Artificial Intelligence Virtual Institute (HAVI) of the US-based Schmidt Sciences. Her interdisciplinary project "Latinitas: Three Millennia of Latin for All" was one of only 19 proposals selected from more than 500 submissions from 42 countries and received the maximum award. A total of 10 million USD was awarded in the funding call. Schmidt Sciences was founded by former Google CEO Eric Schmidt and his wife Wendy Schmidt as a non-profit organization.

Dolganov: “Schmidt Sciences is a global pioneer in promoting innovative humanities research at the intersection of cutting-edge technology. It is a great honor to be part of this initiative.”

Dolganov submitted the proposal together with her research partners David Smith from Northeastern University (Boston, USA), Dimitris Vlitas (SAIL Reply) and Martin Korenjak from the University of Innsbruck. The project will develop the new language model with European AI companies Mistral AI and SAIL Reply. It builds on experience gained from Apollo, the recently unveiled AI model for Ancient Greek, while extending the approach to a significantly larger textual corpus: approximately 12 billion words of Latin are available in digital form, compared with around 600 million words of Ancient Greek. The project aims to combine advanced AI development with the source- and text-critical methods of the Humanities.

The Latin LLM will be integrated into Apollo for Ancient Greek, creating a bilingual AI system that opens up exciting new possibilities for cross-linguistic research. Among other capabilities, it will be able to reconstruct fragmentary texts, make Latin sources semantically and thematically searchable, and retrieve knowledge from extensive textual collections. The project is scheduled to run for two years. 

Dolganov: "By harnessing cutting-edge AI, we aim to open new ways of engaging with Latin – the language of one of the world's largest historical and intellectual traditions. Its written heritage encompasses the literature of antiquity and the Middle Ages, as well as the science and scholarship of early modern Europe, yet much of this vast body of knowledge has remained largely inaccessible.” 

 

Schmidt Sciences 

Schmidt Sciences is a nonprofit organization founded in 2024 by Eric and Wendy Schmidt that works to accelerate scientific knowledge and breakthroughs with the most promising, advanced tools to support a thriving planet. The organization prioritizes research in areas poised for impact including AI and advanced computing, astrophysics, biosciences, climate, and space—as well as supporting researchers in a variety of disciplines through its science systems program.

HAVI 

The Humanities and AI Virtual Institute (HAVI) is a program of Schmidt Sciences that brings together researchers from the humanities and AI to advance both fields. HAVI supports interdisciplinary projects that explore human history, culture and expression while openly sharing the tools and resources they develop, including code, model weights and benchmark datasets.