Question · Languages and accents
How accurate is speech recognition in Lithuanian?
On our Mac benchmark, Lithuanian read speech is transcribed with a word error rate of 17.75 percent by Parakeet and 25.03 percent by Whisper large-v3-turbo. That ranks Lithuanian twenty-third of the 25 languages tested.
Updated October 19, 20263 min read10 cited facts, each traced to its source
Lithuanian is hard for on-device speech recognition, and the engine you pick matters a lot. In our October 2026 benchmark on a Mac with an M2 Pro, Parakeet TDT 0.6B v3 had a word error rate of 17.75 percent 1, about 1 word in 6. Whisper large-v3-turbo had 25.03 percent 2, about 1 word in 4. Both were scored on 100 Lithuanian recordings from the FLEURS test set. We have no Apple SpeechTranscriber result for Lithuanian.
Parakeet is much more accurate than Whisper in Lithuanian
The gap between the engines is large, far beyond what 100 recordings could produce by chance. Parakeet also runs faster: about 20 times faster than real time 3, against about 12 times for Whisper 4. In Lithuanian it is both better and quicker. The speed dataset covers every language.
Our measurement of the engines in Lithuanian
Lower is better in this table.
| Engine | Word error rate (WER %) | Source |
|---|---|---|
| Parakeet TDT v3 | 17.75 | Calculated 1 |
| Whisper large-v3-turbo | 25.03 | Calculated 2 |
The transcripts behind each row are on the on-device benchmark page.
Published results agree on the ranking
NVIDIA reports 20.35 percent for Parakeet on FLEURS Lithuanian 5, a little above our 17.75 percent. OpenAI's paper gives 28.1 percent for the older Whisper large-v2 6, not far from what we measured for the newer Whisper turbo model. Whisper stays well behind Parakeet in every set that has both.
| System | Kind | Test set | Speech | Word error rate (WER %) | Source |
|---|---|---|---|---|---|
| Parakeet TDT 0.6B v3 | Open model | FLEURS | Clean read speech | 17.75 | Calculated 1 |
| NVIDIA Parakeet TDT 0.6B v3 | Open model | FLEURS | Clean read speech | 20.35 | NVIDIA, 2025 5 |
| Whisper large-v3-turbo | Open model | FLEURS | Clean read speech | 25.03 | Calculated 2 |
| Whisper large-v2 | Open model | FLEURS | Clean read speech | 28.1 | Radford et al., 2022 6 |
All published rows are on the published accuracy page.
Lithuanian ranks twenty-third of 25 languages
Taking the best engine in each language, Lithuanian is twenty-third of the 25 languages in the benchmark. Only Slovenian and Maltese are worse. Latvian is just ahead at 17.27 percent 7, and Polish, a larger Slavic neighbor, is far better at 7.70 percent 8. If you only look at Whisper, Lithuanian looks even worse, so the ranking depends on having the right engine.
Formatting makes the text a little worse
Word error rate ignores punctuation and capital letters. Counting them as a reader sees them, Parakeet's text differed from the written sentences by 20.07 errors per 100 tokens 9, and Whisper's by 26.71 10. With Whisper, more than a quarter of tokens are wrong once formatting counts.
What to expect when you dictate in Lithuanian
With Parakeet, expect to correct about 1 word in 6 even in a quiet room. With Whisper, expect closer to 1 in 4, which is slow to fix. Dictation is a rough first draft that you edit. Choose Parakeet if your tool lets you. Names, numbers and inflected endings are the likely trouble spots, and a wrong ending counts as a whole wrong word. Speak in short phrases and proofread. We found no published fact that confirms Lithuanian in a built-in Mac or Windows dictation list, so check your system's language list, and read our overview of dictation languages. If accuracy matters, test a tool on a page of your own text first.
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 specialist vocabulary are not in the test, and we did not measure Apple's model or any cloud service ourselves. Real dictation will usually score worse than read speech. 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 Latvian, Estonian and Polish.
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
The data behind these numbers
- Speech Recognition Accuracy: Word Error Rates6 of this page's figures, 673 data points in all
- On-Device Speech Recognition Benchmark: 25 Languages4 of this page's figures, 56 data points in all
- Speech Recognition Speed: Real-Time Factor Data2 of this page's figures, 66 data points in all
- Dictation Punctuation and Capitalization Accuracy2 of this page's figures, 56 data points in all
Go deeper
Related questions
- Does dictation work for both Brazilian and European Portuguese?
- Does speech recognition understand Swiss German?
- How accurate is speech recognition in Bulgarian?
- How accurate is speech recognition in Croatian?
- How accurate is speech recognition in Czech?
- How accurate is speech recognition in Danish?
- How accurate is speech recognition in Dutch?
- How accurate is speech recognition in English?