Ministry of Digital Transformation of Ukraine announced about the start of the development of a national large language model (LLM), which will be trained on a unique set of Ukrainian data in order to reflect the Ukrainian context as accurately as possible.
It is planned that the first version of this model will be presented by the end of 2025 and will become available to government bodies, educational and scientific institutions, the public sector and business.
To implement the project, the Ministry of Digitization signed a memorandum of cooperation with the mobile operator “Kyivstar”, which will act as a strategic partner and investor. Ukrainian LLM should become a technological basis for the creation of new AI products, development of startups, investment attraction and general growth of the national economy.
The domestic model will better understand the Ukrainian language, culture, and historical context and will be significantly more accurate than foreign counterparts. Having its own language model is also seen as an important step towards the country’s digital independence and a way to preserve critical data within the state.
The model will not only be more accurate in recognizing the Ukrainian context, but also more accessible and cost-effective for Ukrainian users. Its main advantage will be safety: all data will be stored in Ukraine and will be protected taking into account the needs of the state and the defense sphere.
