elevenlabs/tts-v3

Eleven v3 reads text aloud with expressive, emotionally aware delivery across dozens of languages.

$0.10 / 1k characters
GPU: A100

Model Inputs

string

The text to convert to speech

string

The voice to use for speech generation

number

Voice stability (0-1)

boolean

Whether to return timestamps for each word in the generated speech

string

Language code (ISO 639-1) used to enforce a language for the model.

string

This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped.

Result

Idle
Waiting for your input...

Pricing

elevenlabs/

tts-v3

You are billed for what this model actually produces — by the 1k characters — so a shorter request costs less and you never pay for idle time.

Rate
$0.10 / 1k characters
or
10 1k characters / $1

For example, a request of 10 1k characters costs about $1.00.

You are charged for what the request actually produces, so the final amount scales with the 1k characters you use.

ProviderPrice ($)Saving (%)
Synexa$0.05-
fal$0.05same price

Comparison prices are for an equivalent request.

Readme

Eleven v3 is the most expressive model in the Eleven text-to-speech family. It reads punctuation and context to vary pace and emotion rather than delivering a flat monotone.

stability trades consistency against expressiveness: low values give a more dramatic, variable read, high values a steadier one. Set language_code (ISO 639-1) to force a language when the text is ambiguous, and enable timestamps to get per-word timings back alongside the audio.