whatmodel.sh / tts / Fish Speech 1.2
Fish Speech 1.2
↗ share
+ contribute
Fish Speech 1.2 is an open-weights text-to-speech model from Fish Audio, released July 2024. Licence: cc-by-nc-sa-4.0.
Catalogue identity notes
fishaudio/fish-speech-1.2 (base 1.2 release, not 1.2-SFT, 1.4 or 1.5).
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 Fish Speech 1.2, out of 148 with a score, then more from Fish Audio.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 86 |  Chatterbox | 50±5 | api $25.00/M chars +local | ||
| 87 |  Parler-TTS Mini v0.1 | 49±14 | local | ||
| 88 | Fish Speech 1.2 | 49±14 | local | ||
| 89 |  SILMA TTS v1 | 48±10 | local | ||
| 90 |  FireRedTTS (v1) | 48±3 | local | ||
| more from Fish Audio | |||||
| 5 | OpenAudio S1 | 71±8 | api $15.00/M chars | ||
| 31 | Fish Audio S2.1 Pro | 63±3 | api $15.00/M chars | ||
| 37 | Fish Audio S2 Pro | 63±1 | api $15.00/M chars +local | ||
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
other tests1 not in a task · 17% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| TTSDS v1 · default four-factor score | 85.5% ✎ | ▂▂····▃·▃▅▂█ beats 7/14 | 17% |
$ where to run it
Other ways to run it
- Run it as published · needs ~1.15 GB, NVIDIA only · from github.com/fishaudio (official)
git clone https://github.com/fishaudio/fish-speech && cd fish-speech && git checkout 5dd593fef32546f2c0a7d99babbbb27a9478d759 && pip3 install torch torchvision torchaudio && pip3 install -e . && huggingface-cli download fishaudio/fish-speech-1.2 --local-dir checkpoints/fish-speech-1.2 && python tools/llama/generate.py --text "Hello, world." --checkpoint-path "checkpoints/fish-speech-1.2" --num-samples 2 --compile && python tools/vqgan/inference.py -i "codes_0.npy" --checkpoint-path "checkpoints/fish-speech-1.2/firefly-gan-vq-fsq-4x1024-42hz-generator.pth"
Official weights and documentation: huggingface.co/fishaudio ↗
$ how the score was worked out · TTSDS v1 published results 17%
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
| source | results | share of the score |
|---|---|---|
 TTSDS v1 published results | 1 | 17% |
| prior · pull toward the middle | 83% | |
| weighted average | = 48.7 |
$ improve this page
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