Try free

Measured 2026-09-17 on FLEURS

Pikka Speech vs Google Chirp 3: which is more accurate?

Across 77 languages, Pikka Speech records a median character error rate of 2.8% against Google Chirp 3’s 12.1% on the same audio — lower is better. Pikka Speech wins or ties 71 of 77.

Per language

Character error rate, language by language

Lower is better. Difference is Pikka Speech minus Google Chirp 3; a negative value means Pikka Speech is ahead.

Character error rate per language for Pikka Speech and Google Chirp 3 on FLEURS audio.
LanguagePikka SpeechGoogle Chirp 3Difference
Afrikaans2.8%10.7%-7.9 pts
Amharic3.9%12.3%-8.4 pts
Arabic4.1%11.5%-7.4 pts
Armenian12.3%21.1%-8.8 pts
Assamese8.5%42.7%-34.2 pts
Asturian7.6%10.8%-3.2 pts
Azerbaijani0.9%28.0%-27.1 pts
Bengali0.9%10.2%-9.3 pts
Bulgarian0.4%8.2%-7.8 pts
Burmese9.9%7.9%+2.0 pts
Cantonese39.6%12.7%+26.9 pts
Catalan2.2%3.6%-1.4 pts
Chinese4.9%12.8%-7.9 pts
Croatian0.4%2.7%-2.3 pts
Czech6.0%9.1%-3.1 pts
Danish1.8%2.2%-0.4 pts
Dutch1.8%14.2%-12.4 pts
English3.1%9.7%-6.6 pts
Estonian0.5%9.2%-8.7 pts
Finnish0.3%8.8%-8.5 pts
French3.9%16.8%-12.9 pts
Galician1.1%6.1%-5.0 pts
Georgian4.1%4.5%-0.4 pts
German5.1%5.7%-0.6 pts
Greek2.4%9.6%-7.2 pts
Gujarati3.4%15.6%-12.2 pts
Hausa5.9%7.4%-1.5 pts
Hebrew6.7%27.0%-20.3 pts
Hindi1.7%17.3%-15.6 pts
Hungarian4.0%12.1%-8.1 pts
Icelandic10.9%12.1%-1.2 pts
Indonesian1.1%14.5%-13.4 pts
Italian0.3%0.8%-0.5 pts
Japanese1.8%0.9%+0.9 pts
Javanese3.1%6.0%-2.9 pts
Kannada1.6%11.8%-10.2 pts
Kazakh0.9%15.1%-14.2 pts
Khmer9.6%14.3%-4.7 pts
Korean2.2%8.3%-6.1 pts
Kyrgyz5.6%31.0%-25.4 pts
Lao17.6%10.1%+7.5 pts
Latvian2.4%13.9%-11.5 pts
Lithuanian2.1%17.9%-15.8 pts
Luxembourgish35.5%36.1%-0.6 pts
Macedonian1.0%9.3%-8.3 pts
Malay2.2%22.2%-20.0 pts
Malayalam1.2%11.8%-10.6 pts
Maltese12.0%12.9%-0.9 pts
Marathi3.7%26.3%-22.6 pts
Mongolian5.2%5.9%-0.7 pts
Nepali0.0%45.3%-45.3 pts
Northern Sotho13.2%29.2%-16.0 pts
Norwegian0.7%4.2%-3.5 pts
Odia7.8%17.0%-9.2 pts
Persian1.9%3.2%-1.3 pts
Polish1.6%12.1%-10.5 pts
Portuguese1.0%5.4%-4.4 pts
Punjabi4.6%22.1%-17.5 pts
Romanian2.9%60.1%-57.2 pts
Russian0.2%3.2%-3.0 pts
Serbian15.1%81.3%-66.2 pts
Slovak1.2%5.0%-3.8 pts
Slovenian2.0%23.5%-21.5 pts
Spanish1.2%2.1%-0.9 pts
Swahili1.9%7.8%-5.9 pts
Swedish4.5%4.0%+0.5 pts
Tagalog2.3%8.0%-5.7 pts
Tamil16.8%28.8%-12.0 pts
Telugu1.4%23.0%-21.6 pts
Thai4.0%29.5%-25.5 pts
Turkish0.7%44.9%-44.2 pts
Ukrainian1.2%14.7%-13.5 pts
Urdu4.7%29.2%-24.5 pts
Uzbek7.7%14.1%-6.4 pts
Vietnamese1.1%12.2%-11.1 pts
Welsh5.7%25.8%-20.1 pts
Xhosa49.0%44.8%+4.2 pts

Both sides

Where Google Chirp 3 leads

6 of 77 languages, listed in full — the benchmark publishes its losses, not only its wins.

  • Cantonese: Google Chirp 3 12.7% vs Pikka Speech 39.6%
  • Lao: Google Chirp 3 10.1% vs Pikka Speech 17.6%
  • Xhosa: Google Chirp 3 44.8% vs Pikka Speech 49.0%
  • Burmese: Google Chirp 3 7.9% vs Pikka Speech 9.9%
  • Japanese: Google Chirp 3 0.9% vs Pikka Speech 1.8%
  • Swedish: Google Chirp 3 4.0% vs Pikka Speech 4.5%

FAQ

Google Chirp 3 vs Pikka Speech

How is Pikka Speech compared against Google Chirp 3?

Both engines transcribed the same FLEURS audio — 10 utterances per language — through their own production interfaces, and results are scored as character error rate (CER) after normalising case and punctuation. Lower is better. Full per-language detail is on the speech accuracy page.

Where does Google Chirp 3 win?

Google Chirp 3 records a lower error rate on 6 of the 77 languages compared — Cantonese (12.7% vs 39.6%), Lao (10.1% vs 17.6%), Xhosa (44.8% vs 49.0%), Burmese (7.9% vs 9.9%), Japanese (0.9% vs 1.8%). Those rows are published in full on this page.

What does a difference in character error rate mean in practice?

CER is the share of characters an engine gets wrong: 3% means roughly three characters in every hundred differ from what was actually said. In live captions and interpretation, those errors concentrate in names, brands and technical terms — which is why the measured difference matters most on event audio.

Keep exploring

The full benchmark across all engines: speech recognition accuracy. Which languages we cover: language coverage. Or start a free test room.

Hear the difference on your own event

Create a free test room, speak a few sentences, and compare the live captions against Google Chirp 3 on your own audio.