whatmodel.sh / tts / Bark

Bark

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Bark is an open-weights text-to-speech model from Suno, released April 2023. Its published size is 0.9B parameters. Licence: mit.

Catalogue identity notes

suno/bark full-size models (text/coarse/fine). The small models (SUNO_USE_SMALL_MODELS) are a different checkpoint and are not mapped here.

Model information sources

· Parameter count ↗

Official suno/bark card. Value is OUR SUM of the three full-size transformers (3 x 300M); "80/300 M" is read as small/full (the small models are the SUNO_USE_SMALL_MODELS checkpoints, not mapped here). Excludes Meta EnCodec, which Bark uses as its codec. The older GitHub model-card.md lists only 80 M per model.

25.4±2.9#112 of 118 text-to-speech modelsfrom 7 benchmark results across 5 sources

Score within the Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot comparison set only.

$ how it compares

The text-to-speech models ranked either side of Bark, out of 148 with a score.

#modelscorewhereRAMreleased
137Piper (tr_TR dfki medium) Open Home Foundation (Piper)26±4local~0.1GB2023-06
138OpenVoice V2 MyShell26±3local~0.8GB2024-04
139Bark Suno25±3local~4.5GB2023-04
140SpeechT5 TTS Microsoft23±3local~0.7GB2023-02
141MetaVoice 1B v0.1 MetaVoice20±3local~9.7GB2024-02

see all text-to-speech models · find the ones that fit your machine

$ what it's good at

other tests7 not in a task · 97% of its score
testresultvs the others in its rundifficultycountswhy
TTSDS2 listening test · speaker-similarity MOS (4 English domains)2.51 / 5 ✎▅·█▅▅·█▃▃▅·▅ beats 2/17easy · avg 4.03 / 5▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 63%at its level
EmergentTTS · v1 paper automatic MOS (WV-MOS)3.61 / 5 ✎▄·█·█▄····█▄ beats 4/8easy · avg 3.93 / 5▇▇▇▇ 9.1%blunt · at its level
EmergentTTS · 6-category preference9.0% ✎▅▅▃▃█▅▅▃·▃·▅ beats 1/16hard · avg 30%▇▇▇ 7.5%too hard for it
EmergentTTS · 6-category WER20.7% ✎▂▄··▃▄▄··█▄▆ beats 3/16avg 16%▇▇ 6.1%blunt
TTSDS2 listening test · naturalness MOS (4 English domains)2.49 / 5 ✎▅··▃▅··▅█▅██ beats 3/17medium · avg 3.30 / 5▇▇ 5.4%blunt · at its level
TTSDS2 multilingual benchmark · English overall (YouTube, 2025 round)82.2% ✎▂▂█·▂·█▆▂·▄▂ beats 3/17very easy · avg 89%▇▇ 4.7%blunt
TTSDS v1 · default four-factor score73.1% ✎▂▃····▃·▃▅▂█ beats 1/14very easy · avg 86%▇ 0.8%blunt

$ where to run it

Other ways to run it

version

  1. Run it as published · needs ~12 GB · from github.com/suno-ai
    pip install git+https://github.com/suno-ai/bark.git && python -m bark --text "Hello, my name is Suno." --output_filename example.wav

Official weights and documentation: huggingface.co/suno ↗

$ how the score was worked out · TTSDS2 listening test results… 68%, EmergentTTS-Eval six-category… 14%, EmergentTTS-Eval v1 paper (ar… 9% and 2 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
TTSDS2 listening test results (ttsds/listening_test)2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 68%MIT
EmergentTTS-Eval six-category leaderboard2▇▇▇▇ 14%Apache-2.0
EmergentTTS-Eval v1 paper (arXiv:2505.23009v1) · Table 1 automatic MOS1▇▇▇ 9%CC BY 4.0
TTSDS2 multilingual benchmark results (ttsds/leaderboard_data)1▇ 5%MIT
TTSDS v1 published results1▇ 1%MIT
prior · pull toward the middle▇ 3%
weighted average= 25.4

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json