CFStringTokenizerCopyBestStringLanguage showed the best performance for language determination
CFStringTokenizer was released with iOS 3.0 and still is the best option to detect right-to-left language fast.
That's what written in Discussion section in documentation:
The result is not guaranteed to be accurate. Typically, the function requires 200-400 characters to reliably guess the language of a string. CFStringTokenizer recognizes the following languages: ar (Arabic), bg (Bulgarian), cs (Czech), da (Danish), de (German), el (Greek), en (English), es (Spanish), fi (Finnish), fr (French), he (Hebrew), hr (Croatian), hu (Hungarian), is (Icelandic), it (Italian), ja (Japanese), ko (Korean), nb (Norwegian Bokmål), nl (Dutch), pl (Polish), pt (Portuguese), ro (Romanian), ru (Russian), sk (Slovak), sv (Swedish), th (Thai), tr (Turkish), uk (Ukrainian), zh-Hans (Simplified Chinese), zh-Hant (Traditional Chinese).
NSLinguisticTagger was implemented with iOS 5.0 and deprecated in iOS 14.0. It returned the same language for all strings in tests.
Works much longer.
NLLanguageRecognizer was released in iOS 12.0. It is a recommended way to determine language as Apple says, but it is slower even than NSLinguisticTagger.
And, surprise-surprise, 🙃 I've found no case when it determines Persian correctly. It returns Arabic or Urdu. Both solutions mentioned above worked better for Persian.
Maldivian language is not determined by all three ways.
Interface allows to return nil language when it was not determined.
All three ways can return undocumented und language - that also means that language was not detected. Two ways to say the same. Oh, Apple.