Best AI for text to speech
Blind pairwise listening votes. Models still marked preliminary upstream are left out until their rating settles.
Aldena runs these models inside your team's projects. See what each one costs.
| rank | model | vendor | composite | benchmarks | elo |
|---|---|---|---|---|---|
| 1 | Luck Dolphinluck-dolphin/luck-dolphin | Luck Dolphin | 66.7 | 1 of 1 benchmarks | 1684 |
| 2 | Inworld TTS (max)inworld/inworld:max | Inworld | 65.1 | 1 of 1 benchmarks | 1567 |
| 3 | Inworld TTS 1.5 MAXinworld/inworld-max-1.5 | Inworld | 63.6 | 1 of 1 benchmarks | 1565 |
| 4 | CastleFlow v1.0async/async-1 | Async | 62.1 | 1 of 1 benchmarks | 1557 |
| 5 | Hume Octavehume/hume-octave | Hume | 59.9 | 1 of 1 benchmarks | 1551 |
| 6 | Papla P1papla/papla-p1 | Papla | 59.9 | 1 of 1 benchmarks | 1551 |
| 7 | Inworld TTSinworld/inworld | Inworld | 57.6 | 1 of 1 benchmarks | 1543 |
| 8 | OpenAudio S2lanternfish/lanternfish-2 | Lanternfish | 56.1 | 1 of 1 benchmarks | 1542 |
| 9 | Eleven Multilingual v2elevenlabs/eleven-multilingual-v2 | ElevenLabs | 54.5 | 1 of 1 benchmarks | 1507 |
| 10 | Eleven Turbo v2.5elevenlabs/eleven-turbo-v2.5 | ElevenLabs | 53.0 | 1 of 1 benchmarks | 1504 |
| 11 | Eleven Flash v2.5elevenlabs/eleven-flash-v2.5 | ElevenLabs | 50.8 | 1 of 1 benchmarks | 1503 |
| 12 | MiniMax Speech 02 HDminimax/minimax-speech-02-hd | MiniMax | 50.8 | 1 of 1 benchmarks | 1503 |
| 13 | MiniMax Speech 02 Turbominimax/minimax-speech-02-turbo | MiniMax | 48.5 | 1 of 1 benchmarks | 1482 |
| 14 | Typecast SSFM 3.0typecast/typecast | Typecast | 47.0 | 1 of 1 benchmarks | 1475 |
| 15 | Kokoro v1.0hexgrad/kokoro-v1 | Hexgrad | 45.5 | 1 of 1 benchmarks | 1469 |
| 16 | OpenAudio S1lanternfish/lanternfish-1 | Lanternfish | 43.9 | 1 of 1 benchmarks | 1465 |
| 17 | Chatterboxresemble-ai/chatterbox | Resemble AI | 42.4 | 1 of 1 benchmarks | 1460 |
| 18 | Magpie Research Previewmagpie/magpie-rp | Magpie | 40.9 | 1 of 1 benchmarks | 1456 |
| 19 | Hithink Speech 2.6hithink/hithink-speech-2.6 | Hithink | 39.4 | 1 of 1 benchmarks | 1454 |
| 20 | NeuTTS Maxneuphonic/neuphonic | Neuphonic | 37.9 | 1 of 1 benchmarks | 1389 |
| 21 | Maya 1maya-research/maya1 | Maya Research | 36.4 | 1 of 1 benchmarks | 1383 |
| 22 | Magpie Multilingualmagpie/magpie | Magpie | 34.8 | 1 of 1 benchmarks | 1372 |
| 23 | Wordcab TTSwordcab/wordcab | Wordcab | 33.3 | 1 of 1 benchmarks | 1331 |
How this ranks
Every benchmark value becomes a percentile among the models that have it, so accuracy scores, Elo ratings and word error rates compare without hand-tuned scaling. Metrics where lower is better are inverted first. Raw values are never summed or averaged across benchmarks. A model's mean percentile is then shrunk toward the mean of the models that were broadly benchmarked, so a model tested twice cannot outrank a broadly tested one on two lucky results. Turning a data source off runs that same ranking code again in your browser over the sources you left on.
A model scored on fewer than 1 of the 1 ranked benchmarks in this category still ranks here, on the benchmarks it does have, and its row carries a partial coverage mark. On an equal score it sits under the model that earned the same number across more of the board.
Data sources
Turn a source off to drop every benchmark it feeds and re-rank the board in your browser. Your choice follows you across the leaderboard pages.
- TTS Arena V2
Elo ratings from TTS Arena V2.