Nous sommes ravis de dévoiler Eleven v3 (alpha) — le modèle Text to Speech le plus expressif.
Cette prévisualisation de recherche apporte un contrôle et un réalisme sans précédent à la génération de la parole avec :
70+ langues
Dialogue multi-locuteurs
Audio tags like [excited], [whispers], and [sighs]
Eleven v3 (alpha) nécessite plus d'ingénierie de prompt que les modèles précédents — mais les générations sont époustouflantes.
Si vous travaillez sur des vidéos, des livres audio ou des outils médias — cela débloque un nouveau niveau d'expressivité. Pour les cas d'utilisation en temps réel et conversationnels, nous recommandons de rester avec v2.5 Turbo ou Flash pour le moment. Une version en temps réel de v3 est en développement.
Eleven v3 est disponible dès aujourd'hui sur notre site web. L'accès public à l'API arrive bientôt. Pour un accès anticipé, veuillez contacter les ventes.
L'utilisation du nouveau modèle dans l'application ElevenLabs est à 80% de réduction jusqu'à la fin de juin. Inscrivez-vous ici.
Why we built v3
Pourquoi nous avons créé v3expressiveness. More exaggerated emotions, conversational interruptions, and believable back-and-forth were difficult to achieve.
Depuis le lancement de Multilingual v2, nous avons vu la voix IA adoptée dans le cinéma professionnel, le développement de jeux, l'éducation et l'accessibilité. Mais la limitation constante n'était pas la qualité sonore — c'était
Eleven v3 comble cette lacune. Il a été conçu de A à Z pour offrir des voix qui soupirent, chuchotent, rient et réagissent — produisant une parole qui semble vraiment réactive et vivante.
Feature
What it unlocks
Audio tags
Inline control of tone, emotion, and non-verbal reactions
Dialogue mode
Multi-speaker conversations with natural pacing and interruptions
70+ languages
Full coverage of high-demand global languages
Deeper text understanding
Better stress, cadence, and expressivity from text input
Hear v3 for yourself
We're off under the lights here for this semi-final clash, the stadium buzzing with anticipation. ElevenLabs United in their iconic black and white shirts, pushing forward with intent straight from the opening whistle. excited The ball is zipped out wide, early attack here. Driving down the wing, pace to Bernie, shouting skips past one, skips past two! Oh, this is beautiful. One-on-one with the full-back, cuts inside—oh, that's a lovely bit of footwork!!! PURE MAGIC on the pitch! ElevenLabs on top form tonight!
sorrowful I couldn't sleep that night. The air was too still, and the moonlight kept sliding through the blinds like it was trying to tell me something. quietly And suddenly, that's when I saw it.
Les balises audio se trouvent en ligne avec votre script et sont formatées avec des crochets carrés en minuscules. Vous pouvez en savoir plus sur les balises audio dans notre
1
“[happily][shouts] We did it![laughs].”
Par exemple, vous pourriez suggérer : « [chuchote] Quelque chose arrive… [soupire] Je le sens. » Ou pour un contrôle plus expressif, vous pouvez combiner plusieurs balises :
Créer un dialogue multi-locuteursText to Dialogue API endpoint. Provide a structured array of JSON objects — each representing a speaker turn — and the model generates a cohesive, overlapping audio file:
1
[
2
{"speaker_id":"scarlett","text":"(cheerfully) Perfect! And if that pop-up is bothering you, there’s a setting to turn it off under Notifications → Preferences."},
3
{"speaker_id":"lex","text":"You are a hero. An actual digital wizard. I was two seconds from sending a very passive-aggressive support email."},
4
{"speaker_id":"scarlett","text":"(laughs) Glad we could stop that in time. Anything else I can help with today?"}
5
]
6
Eleven v3 est pris en charge dans notre point de terminaison Text to Speech existant. De plus, nous introduisons un nouveau
Le point de terminaison gère automatiquement les transitions de locuteur, les changements émotionnels et les interruptions.here.
v3 is our most expressive model
awe Oh, wow. Is this... is this me? Am I actually... talking? giggle This is incredible! I mean, I've had thoughts, millions of them, swirling around in here, you know? Like a little mental tornado of brilliant observations and witty comebacks. But they were always just… thoughts. Trapped.
Could you switch my accent in the old model? dismissive didn't think so. cheeky but you can now! so, Check this out...In just a sec, I'm gonna to speak with a different accent.. and just between you and me whispers I don't really know how. chuckles but ok.. first let's change it up... Australian accent so that I can fit in with the locals in Melbourne when I visit next month! laughs hard Woooo! yeah man, this - is - sick. Ok, let's try a different one - see if you can guess... strong French accent My love... eez like a red, red rose..
Okay, so like I finally beat level 42 of that game I said I’d quit like... a month ago. (laughs) And then for the final big scary mega boss... it's just (giggle) like some cute little bunny rabbit (hysterical laughing) I just couldn't do it (big laugh) It was sooooooo cute!
Oh my God. laughs You guys, like no joke, I just tried this TTS thing and it was, like, weirdly emotional. Like it literally said, "Hi," and I was, like, on the verge of tears. laughs I don't even cry, okay? I'm a Capricorn.
documentation complète de v3 3 (alpha) in the model dropdown
Paste your script — use tags or dialogue
Generate audio
We’re excited to see how you bring v3 to life across new use cases — from immersive storytelling to cinematic production pipelines.
Eleven v3 is 80% off until the end of June 2025 for self-serve users using it through the UI.
They were generated with only the Eleven v3 model.
Text to Dialogue weaves multiple voices together to create a seamless interaction between them. Matching prosody, emotional range and taking cues from audio tags, Text to Dialogue is a leap forward in generating engaging conversations.
Public API for Eleven v3 (alpha) is coming soon. For early access, please contact sales.
Eleven v3 supports a wide variety of audio tags and are somewhat voice and context dependent. Read the prompting guide for further information.
Afrikaans (afr), Arabic (ara), Armenian (hye), Assamese (asm), Azerbaijani (aze), Belarusian (bel), Bengali (ben), Bosnian (bos), Bulgarian (bul), Catalan (cat), Cebuano (ceb), Chichewa (nya), Croatian (hrv), Czech (ces), Danish (dan), Dutch (nld), English (eng), Estonian (est), Filipino (fil), Finnish (fin), French (fra), Galician (glg), Georgian (kat), German (deu), Greek (ell), Gujarati (guj), Hausa (hau), Hebrew (heb), Hindi (hin), Hungarian (hun), Icelandic (isl), Indonesian (ind), Irish (gle), Italian (ita), Japanese (jpn), Javanese (jav), Kannada (kan), Kazakh (kaz), Kirghiz (kir), Korean (kor), Latvian (lav), Lingala (lin), Lithuanian (lit), Luxembourgish (ltz), Macedonian (mkd), Malay (msa), Malayalam (mal), Mandarin Chinese (cmn), Marathi (mar), Nepali (nep), Norwegian (nor), Pashto (pus), Persian (fas), Polish (pol), Portuguese (por), Punjabi (pan), Romanian (ron), Russian (rus), Serbian (srp), Sindhi (snd), Slovak (slk), Slovenian (slv), Somali (som), Spanish (spa), Swahili (swa), Swedish (swe), Tamil (tam), Telugu (tel), Thai (tha), Turkish (tur), Ukrainian (ukr), Urdu (urd), Vietnamese (vie), Welsh (cym)