Voice · ElevenLabs
ElevenLabs
ElevenLabs' voice models, from the fast Turbo v2.5 to the expressive Eleven v4. Pick Eleven v4 for natural narration in French or another of its 90+ languages (about 12 credits for a 3,000-character page).
Anonymized High costWhat sets it apart
ElevenLabs' voice line runs from Multilingual v2 to the expressive Eleven v3 and v4, with audio tags and voice cloning on the newer ones, and a real-time Turbo variant of v4. ElevenLabs now lists the older Turbo v2.5 as deprecated in favor of Flash v2.5.
- Up to 90+ languages, French included
- Fast: Flash v2.5 reaches about 75 ms
- v3 and v4 add audio tags; v4 Turbo is real time
Recommended for: read a note or narrate a film.
Strengths and limits
Where it shines
- 29 to 90+ languages depending on the model, French included
- v4 Turbo is built for real-time voice agents
Think twice when
- About 6 to 16 credits for a 3,000-character page, depending on the model
- Anonymized, not private
- 21 voices
- Up to 5,000 characters per request in Sub Rosa
Use cases
Narrated product video in French
Narrate this product description in warm, clear French for a 30 second ad.
Eleven v4 covers 90+ languages with expressive voices.
Voice replies in an app
Say: Your appointment is confirmed for Tuesday at 3 pm.
v4 Turbo is built for real-time voice agents.
Against its rivals
- Gemini 3.1 Flash TTSGemini Flash TTS also steers delivery with tags; Eleven v4 scores higher on the Speech Arena than the Gemini 3.8 Flash TTS entry.#3 of 10 on Artificial Analysis Speech Arena (1,209 Elo)
- MiniMax Speech-02 HDMiniMax Speech-02 HD has a close arena score and covers 30+ languages.#6 of 10 on Artificial Analysis Speech Arena (1,114 Elo)
- Inworld TTS-1.5 MaxInworld costs about a fifth as much with 15 languages.
Version history
Each version, dated, with what it changed over the one before.
Hollow points are earlier versions no longer offered in Sub Rosa.
See the numbers
| Version | Released | Elo |
|---|---|---|
| Multilingual v2 | 2023-08 | 1,095 |
| Turbo v2.5 | 2024-07 | 1,097 |
| Flash v2.5 | 2024-12-18 | 1,078 |
| Eleven v3 | 2025-06-03 | 1,178 |
| Eleven v4 Turbo | 2026-09-28 | 1,330 |
| Eleven v4 | 2026-09-28 | 1,329 |
- Multilingual v2
- Emotion-aware multilingual voice, 29 languages, stable on long-form content
- Turbo v2.5
- Low-latency model covering 32 languages
- Flash v2.5Earlier, not in Sub Rosa
- Faster still: speech in about 75 ms plus network latency, 32 languages
- Eleven v3
- 70+ languages, inline audio tags such as [whispers], multi-speaker dialogue
- Eleven v4
- 90+ languages, better voice cloning from 10 s of audio, inline audio tags, 10,000 characters per request in the API
- Eleven v4 TurboLatest
- Low-latency variant of v4 for voice agents: median inference about 100 ms, bidirectional streaming
- Same 90+ languages, voice cloning from 10 s of audio and inline audio tags as v4
Benchmarks
How to read these scores: what a benchmark tells you.
Artificial Analysis Speech Arena: #1 of 10 models measured in the catalog. People listen to the same sentence read by two voices, without knowing the models, and pick the more natural one. An Elo rating: higher sounds more natural to more people. 100 points apart means the higher-rated model wins about 64% of duels.
- 1. ElevenLabs TTS v4 TurboElevenLabs1,330
- 2. ElevenLabs TTS v4ElevenLabs1,329
- 3. Gemini 3.1 Flash TTSGemini 3.1 Flash TTS1,209
- 4. ElevenLabs TTS v3ElevenLabs1,178
- 5. Gradium TTSGradium TTS1,151
- 6. MiniMax Speech-02 HDMiniMax Speech-02 HD1,114
- 7. Chatterbox HD (Resemble AI)Chatterbox HD1,102
- 10. Kokoro Text to SpeechKokoro1,064
See the numbers
| Model | Elo |
|---|---|
| 1. ElevenLabs TTS v4 Turbo | 1,330 |
| 2. ElevenLabs TTS v4 | 1,329 |
| 3. Gemini 3.1 Flash TTS | 1,209 |
| 4. ElevenLabs TTS v3 | 1,178 |
| 5. Gradium TTS | 1,151 |
| 6. MiniMax Speech-02 HD | 1,114 |
| 7. Chatterbox HD (Resemble AI) | 1,102 |
| 10. Kokoro Text to Speech | 1,064 |
All 5 models in this family
| Model | Use | Privacy | Price |
|---|---|---|---|
ElevenLabs Turbo v2.5tts-elevenlabs-turbo-v2-5 | Voice | Anonymized | $0.0281 per 1,000 characters$0.0843 a page read aloud |
ElevenLabs Multilingual v2elevenlabs-tts-multilingual-v2 | Voice | Anonymized | $0.0517 per 1,000 characters$0.155 a page read aloud |
ElevenLabs TTS v3elevenlabs-tts-v3 | Voice | Anonymized | $0.0517 per 1,000 characters$0.155 a page read aloud |
ElevenLabs TTS v4elevenlabs-tts-v4 | Voice | Anonymized | $0.0414 per 1,000 characters$0.124 a page read aloud |
ElevenLabs TTS v4 Turboelevenlabs-tts-v4-turbo | Voice | Anonymized | $0.0207 per 1,000 characters$0.062 a page read aloud |
Specifications
| Specification | Eleven v4 | Turbo v2.5 | Eleven v4 Turbo | Eleven v3 | Multilingual v2 |
|---|---|---|---|---|---|
| Released | September 28, 2026 | July 2024 | September 28, 2026 | June 3, 2025 | August 2023 |
| Weights | Closed | Closed | Closed | Closed | Closed |
| Accepts | text | text | text | text | text |
| Produces | audio | audio | audio | audio | audio |
| Languages | 90+ languages | 32 languages | 90+ languages | 70+ languages | 29 languages |
| Price | $0.0414 per 1,000 characters | $0.0281 per 1,000 characters | $0.0207 per 1,000 characters | $0.0517 per 1,000 characters | $0.0517 per 1,000 characters |
| Privacy | Anonymized | Anonymized | Anonymized | Anonymized | Anonymized |
Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.
Sources
Everything on this page comes from these pages, read on October 16, 2026 or as dated on each score.