Urdu translation and speech-to-text models ur

3 models into Urdu, 2 from Urdu, 12 speech-to-text. Best scored into Urdu: Hindi to Urdu (chrF++ 23.0). Best scored from Urdu: Urdu to English (chrF++ 30.6).

Which open models work for Urdu?

Windy Translate lists 3 open translation models into Urdu and 2 from Urdu. All of them can be downloaded from Hugging Face and run on your own hardware. By licence: 3 Apache-2.0 and 2 CC-BY-4.0 (terms and attribution text: licensing). 3 of them have a screening score on FLORES-200. The best scored into Urdu is translate-hi-ur (chrF++ 23.0, Limited). The best scored from Urdu is translate-ur-en (chrF++ 30.6, Usable).

With English. English to Urdu: translate-en-ur (chrF++ 22.1, Limited), the best of 2 direct models. Urdu to English: translate-ur-en (chrF++ 30.6, Usable), the best of 2 direct models. To translate between Urdu and a language with no direct model, you can go through English with these two.

By band, the scored Urdu models are 1 Usable and 2 Limited (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. Whisper's tokenizer lists Urdu as one of its languages, so all 12 multilingual Whisper-based models here can transcribe it: see the speech models. Accuracy differs a lot between languages and model sizes, and this site has not measured speech accuracy.

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 Urdu

From Urdu

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