How accurate is speech recognition in Slovenian?

Slovenian is one of the hardest languages for on-device speech recognition: about 1 wrong word in 5 on read speech. In our October 2026 benchmark on a Mac with an M2 Pro, Parakeet TDT 0.6B v3 had a word error rate of 19.83 percent ¹ and Whisper large-v3-turbo had 23.51 percent ², on 100 Slovenian recordings from the FLEURS test set. We have no Apple SpeechTranscriber result for Slovenian.

Parakeet is about 4 points better than Whisper in Slovenian

Here the engine choice matters. Parakeet beats Whisper by close to four points of word error rate, a gap that is large relative to the sample noise. Parakeet is also faster: about 20 times faster than real time ³, against about 12 times for Whisper ⁴. In Slovenian it is better and quicker. The speed dataset covers every language.

Our measurement of the engines in Slovenian

Lower is better in this table.

The transcripts behind each row are on the on-device benchmark page.

Published results confirm the weakness and vary by test

NVIDIA reports 24.03 percent for Parakeet on FLEURS Slovenian ⁵, higher than our 19.83 percent, and OpenAI's paper gives 23.1 percent for the older Whisper large-v2 ⁶. On CoVoST 2 Slovenian, Parakeet had 31.80 percent ⁷, much worse. Every published figure is far above those for Spanish or Portuguese, but the exact number depends on the recordings and scoring.

All published rows are on the published accuracy page.

Slovenian ranks twenty-fourth of 25 languages

Taking the best engine in each language, Slovenian is twenty-fourth of the 25 languages in the benchmark. Only Maltese is worse. Latvian (18.64 percent with Whisper ⁸) is just ahead, and Croatian (11.34 percent ⁹), its neighbor, is much better. Portuguese (3.28 percent ¹⁰) shows how far the best languages are ahead. Slovenian is a small language with little training audio, and it has rich word endings, though the benchmark does not test which factor matters more.

What to expect when you dictate in Slovenian

Expect to correct about 1 word in 5 even in a quiet room, so dictation works as a rough first draft that you edit, not as finished text. Use Parakeet where you can, and prefer short, clear phrases. Names, numbers and word endings are the likely trouble spots. If accuracy matters, compare a cloud service on your own text, and read our overview of dictation languages before you choose a tool. We found no published fact that confirms Slovenian in a built-in Mac or Windows dictation list, so check your system's language list.

How to choose an engine for Slovenian

In Slovenian the engine choice is not a toss-up. Parakeet has the lower error and is the faster of the two, so it is the engine to try first. If your tool offers only Whisper, expect to edit more. Because the error rate is high, test any tool on a page of your own text before you commit to dictating long documents with it, and consider typing the passages where exact names and numbers matter. The published accuracy page lists other models and test sets, so you can see whether any of them does clearly better on Slovenian.

What these numbers do not show

The recordings are people reading prepared sentences in a quiet room, scored on one Mac with 100 recordings. Spontaneous speech, dialects, noise and domain vocabulary are not in the test, and we did not measure Apple's model or any cloud service ourselves. The error rate also counts wrong word endings as whole wrong words, which is strict for Slovenian. For the measure itself, see the guide on how accurate speech recognition is.

Our guide to European languages in speech to text helps you choose an engine, and the languages overview covers other languages. For comparison, read the same question for Croatian and Italian.