We standardize language codes before comparing language support or making translation requests.
We convert the code before every call to an MT engine, in 24 of the 100 languages. An LLM gets the name of the language in the prompt.
Language-code conversions before an API call
| Language | Our code | Model codes |
|---|---|---|
| Arabic (MSA) | ar |
|
| Bulgarian | bg |
|
| Cebuano | ceb |
|
| Chinese Simplified | zh |
|
| Chinese Traditional | zh-hant |
|
| Danish | da |
|
| Estonian | et |
|
| Finnish | fi |
|
| French | fr |
|
| German | de |
|
| Italian | it |
|
| Japanese | ja |
|
| Korean | ko |
|
| Mongolian | mn |
|
| Norwegian | no |
|
| Portuguese (European) | pt |
|
| Romanian | ro |
|
| Serbian | sr-latn |
|
| Slovenian | sl |
|
| Spanish (European) | es |
|
| Swedish | sv |
|
| Uzbek | uz |
|
| Vietnamese | vi |
|
| Yiddish | yi |
|