Tasks and inputs
Eleven v4 is listed for Text → Speech, Text-to-speech. It accepts text and produces audio. Use the publisher’s input format and preprocessing for your chosen deployment.
Eleven v4 is ElevenLabs' flagship text-to-speech model for audiobooks, dubbing and emotional dialogue, with 90+ languages and 10,000 characters per request.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: About $0.08 per 1K characters ($80 per 1M characters) at list price; a promotional rate ran to 12 October 2026. Indicative; confirm on your ElevenLabs plan.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Narration is where Eleven v4 earns its keep. It is ElevenLabs' highest-quality, most emotive speech model, released on 28 September 2026, and its design choices all point at long, expressive output: audiobooks, game and animation voiceovers, dubbing, dialogue with real emotional range.
Compared with Eleven v3, the jump is in reach and length. Language coverage goes from 70+ to 90+, and a single request can now hold 10,000 characters, roughly ten minutes of audio, where v3 stopped at 5,000.
Delivery is directed in two ways. Inline audio tags mark emotion, sound effects and accents at specific points in the text. Natural-language direction describes the performance you want more broadly. IPA phoneme support is better too, useful for brand names and invented words in fiction.
Multi-speaker dialogue is native. The model is served through the Text to Dialogue API, so a scene with several characters can be one request instead of stitched clips.
clone_strength setting from 0 to 1 arrived on 26 October.That last point catches people out. If your catalogue relies on older professional clones, budget time to retrain them before switching.
| Field | Value |
|---|---|
| Model ID | eleven_v4 |
| Endpoint | Text to Dialogue API |
| Languages | 90+ (new versus v3 include Amharic, Lao, Kamba and Fula) |
| Max per request | 10,000 characters |
| Indicative price | $0.08 per 1K characters ($80 per 1M) |
| Replaces | eleven_v3 |
A short promotional rate ran until 12 October. The list price above is indicative; your plan may differ.
Launch messaging described v4 as fast, but the model documentation reserves low-latency claims for its real-time sibling, Eleven v4 Turbo, at half the price. If a user is waiting on the other end of a call, Turbo is the right choice. Pick this flagship when nobody is waiting and the performance itself is the product.
Choose three passages from your own material: a quiet narrative stretch, a heated argument between two characters and a paragraph full of proper nouns. Render each with v3 and v4 using the same voice, then have listeners rank them blind. Count retakes per finished minute; that, not the per-character rate, is the real production cost. Compare also with MiniMax Speech 2.8 HD and Gemini 3.8 Flash TTS, or browse all text-to-speech models.
Compare the model on the work you need to complete.
Eleven v4 is listed for Text → Speech, Text-to-speech. It accepts text and produces audio. Use the publisher’s input format and preprocessing for your chosen deployment.
The listed route has starting rates of $80.00 input per million tokens. Compare pronunciation, voice consistency and time to the first playable audio. Check the service’s billing unit for the actual script.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
inclusionAI
Ming-3.0-tts-0.5B is a compact open text-to-speech model from inclusionAI with voice design, zero-shot cloning and fine control over delivery.
Sarvam
Bulbul V4 Flash is Sarvam's fast text-to-speech model for 11 Indian locales, with sub-250 ms first audio at a third less than Bulbul V3.
ElevenLabs
Eleven v4 Turbo is ElevenLabs' low-latency v4 speech model for voice agents, with about 100 ms median inference, 90+ languages and half v4's price.