whatmodel.sh / tts / Gemini 2.5 Flash Preview TTS
Gemini 2.5 Flash Preview TTS
↗ share
+ contribute
Gemini 2.5 Flash Preview TTS is a closed-weights text-to-speech model from Google, released May 2025.
Catalogue identity notes
Gemini-2.5-Flash-Preview-TTS as explicitly linked in the result row.
Model information 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 Gemini 2.5 Flash Preview TTS, out of 148 with a score, then more from Google.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 1 |  Seed TTS 2.0 Standard | 84±5 | comparison only | ||
| 2 |  Azure Neural TTS Xiaoxiao | 74±4 | comparison only | ||
| 3 | Gemini 2.5 Flash Preview TTS | 72±3 | comparison only | ||
| 4 |  Voxtral TTS (4B) | 71±4 | api $16.00/M chars +local | ||
| 5 |  OpenAudio S1 | 71±8 | api $15.00/M chars | ||
| more from Google | |||||
| 6 | Gemini 2.5 Pro Preview TTS | 69±3 | comparison only | ||
| 8 | Gemini 3.1 Flash TTS Preview | 69±2 | api $50.00/M chars | ||
| 22 |  Chirp 3: HD | 65±4 | api $30.00/M chars | ||
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
other tests7 not in a task · 96% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| EmergentTTS · 6-category preference | 75.6% ✎ | ▅▅▃▃█▅▅▃·▃·▅ beats 16/16 | 26% |
| InstructTTSEval ZH · mean instruction-following score (APS, DSD, RP) | 85.4% ✎ | ▄··········█ beats 2/2 | 22% |
| RW-Voice-EQ TTS · Acting / Role-Fit (human) | 3.64 / 5 ✎ | ▄·····█····▄ beats 2/3 | 16% |
| RW-Voice-EQ TTS · Reliability (human) | 3.66 / 5 ✎ | █·▄······▄·▄ beats 2/4 | 10% |
| InstructTTSEval EN · mean instruction-following score (APS, DSD, RP) | 88.7% ✎ | ▄····█··▄··▄ beats 4/4 | 8.9% |
| RW-Voice-EQ TTS · Expressiveness (human) | 3.44 / 5 ✎ | ███···█····█ beats 3/4 | 6.8% |
| EmergentTTS · 6-category WER | 10.4% ✎ | ▂▄··▂▄▄··█▄▆ beats 15/16 | 6.4% |
$ where to run it
No verified API or download recorded yet · add one
$ how the score was worked out · RW-Voice-EQ Bench (Hume AI) 33%, EmergentTTS-Eval six-category… 32%, InstructTTSEval: instruction-… 31%
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 |
|---|---|---|
 RW-Voice-EQ Bench (Hume AI) | 3 | 33% |
 EmergentTTS-Eval six-category leaderboard | 2 | 32% |
| InstructTTSEval: instruction-following TTS benchmark (arXiv:2506.16381v1) | 2 | 31% |
| prior · pull toward the middle | 4% | |
| weighted average | = 71.5 |
$ 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