Serbo-Croatian translation models sh

4 models into Serbo-Croatian, 3 from Serbo-Croatian. Best scored into Serbo-Croatian: Ukrainian to Serbo-Croatian (chrF++ 2.7). Best scored from Serbo-Croatian: Serbo-Croatian to English (chrF++ 62.5).

Which open models work for Serbo-Croatian?

Windy Translate lists 4 open translation models into Serbo-Croatian and 3 from Serbo-Croatian. All of them can be downloaded from Hugging Face and run on your own hardware. By licence: 5 Apache-2.0 and 2 CC-BY-4.0 (terms and attribution text: licensing). 6 of them have a screening score on FLORES-200. The best scored into Serbo-Croatian is translate-uk-sh (chrF++ 2.7, Not recommended). The best scored from Serbo-Croatian is translate-tc-big-sh-en (chrF++ 62.5, Excellent).

With English. English to Serbo-Croatian: translate-tc-base-en-sh (not yet scored). Serbo-Croatian to English: translate-tc-big-sh-en (chrF++ 62.5, Excellent). To translate between Serbo-Croatian and a language with no direct model, you can go through English with these two.

By band, the scored Serbo-Croatian models are 1 Excellent, 1 Usable, 1 Limited and 3 Not recommended (Excellent is chrF++ 60 or more, Good 45 to 60, Usable 30 to 45, Limited 15 to 30). A score is a 48-sentence screening of news and encyclopedia text, not a measure of accuracy, and a difference of a point or two is noise. How we measured.

Speech-to-text. No speech model here covers Serbo-Croatian: there is no dedicated model, and Whisper's tokenizer does not list the language.

Each model page has a short snippet to download and run the model yourself, its licence and attribution text, and its score where there is one. Pick the best scored model for your direction, then try it on a few sentences of your own text: a screening score is a starting point, not a guarantee.

Into Serbo-Croatian

From Serbo-Croatian

Scores: chrF++ on FLORES-200 dev, 48 sentences, measured 2026-08-04. Method.