Eleven v4 Turbo
Eleven v4 Turbo is the low-latency variant of ElevenLabs' Eleven v4, released alongside it on 28 September 2026 for real-time use such as voice agents. ElevenLabs says it keeps v4's expressive range, supports more than 90 languages and reaches a median of about 100 ms inference latency and 150 ms to first speech over WebSocket streaming. It costs $0.04 per thousand characters on fal and on ElevenLabs' API, half of Eleven v4.
Anvisha Pai, Co-founder & CEO, Voyager
ElevenLabs · released
›What we asked for
Hi, I'm Eleven v4 Turbo, the fast version of ElevenLabs' newest voice model. I'm built for voice agents that need to answer in real time. Here is the same script every voice on this site reads, so you can compare us.
elevenlabs/tts/eleven-v4-turbo
voice: Sarah, stability: 0.5
returned 14.0 seconds of mp3
At a glance
- Price
- $0.04 per thousand characters on fal and on ElevenLabs' API, half of Eleven v4; our four scripts cost about four cents.
- Voices
- The same twenty named stock voices as Eleven v4 on fal; we used Sarah.
- Latency
- ElevenLabs quotes about 150 ms to first speech on a stream; whole files from fal took four to seven seconds for us.
- Languages
- More than 90, per ElevenLabs.
Other ElevenLabs models: Eleven v4, Eleven v3. Compare all 3.
Anvisha's take
We have not scored the four readings by ear; press play and judge the voice yourself. What we measured: Turbo read the four scripts in 14 to 26 seconds of audio each and returned in four to seven seconds, for four cents across all four. That is half of Eleven v4's cost, but not much quicker: v4 took four to eight seconds for the same files. fal's endpoint returns a finished file, so the 150 ms to first speech ElevenLabs advertises, which is measured on a stream, is not what this test sees. Our readings ran from 14.0 to 26.4 seconds long, within a second of v4's on every script, with the same Sarah voice and stability.
Eleven v4 Turbo real output: the same 4 scripts every model gets
Every voice model on this site gets the same 4 scripts, so you can compare like with like. These are the readings Eleven v4 Turbo returned, one try each, nothing re-rolled or retouched. Generated on . The 4 readings took 4 to 6 seconds each and cost $0.039 in total.
Pacing, warmth and whether the pauses land where a voice actor would put them.
›What we asked for
Meet the Ember mug. It keeps your coffee at exactly the temperature you choose, for up to ninety minutes, and it tells your phone when it's ready. Warm from the first sip to the last. Ember. Coffee, on your terms.
elevenlabs/tts/eleven-v4-turbo
voice: Sarah, stability: 0.5
returned 14.0 seconds of mp3
Steady narration over a long paragraph, clean list rhythm and a natural close.
›What we asked for
A pour-over works in four steps. First, rinse the paper filter with hot water so it doesn't taste of paper. Second, add the grounds and pour just enough water to wet them, then wait thirty seconds while they bloom. Third, pour the rest of the water in slow circles. Fourth, wait about three minutes for it to drip through. That's it: a cup that tastes like the coffee, not the machine.
elevenlabs/tts/eleven-v4-turbo
voice: Sarah, stability: 0.5
returned 26.4 seconds of mp3
Range: excitement, then a hushed aside, from the words alone with no tags.
›What we asked for
We did it! We actually did it! I can't believe it's finally over. Okay. Okay, deep breath. Let's not tell anyone until Monday, alright? Just... let me have this one quiet night.
elevenlabs/tts/eleven-v4-turbo
voice: Sarah, stability: 0.5
returned 14.2 seconds of mp3
Reading numbers, times, currency, an email address and two hard names correctly.
›What we asked for
Your order ships on the 24th of September 2026 and arrives by 9:15 a.m. The total is $1,249.99, including VAT. Questions? Ask for Siobhan Nguyen or Ravi Parikh on 0800 555 0199, or email help@moda.app.
elevenlabs/tts/eleven-v4-turbo
voice: Sarah, stability: 0.5
returned 22.3 seconds of mp3
How the cost under each picture was worked out: fal's Eleven v4 Turbo rate of $0.04 per 1,000 characters (fal model page and pricing API, read 2026-10-22) for the 213 characters sent; fal does not return the charge.
Compare Eleven v4 Turbo
Eleven v4 Turbo pricing
| Where | Unit | Price | Source |
|---|---|---|---|
| fal | 1,000 characters | $0.04 | fal model page and pricing API, read |
| ElevenLabs API | 1,000 characters The same rate as Flash and Turbo v2.5 there. | $0.04 | ElevenLabs API pricing, read |
Is Eleven v4 Turbo free?
ElevenLabs' announcement invites free accounts to generate with v4 Turbo, and the free plan gives 10,000 credits a month without a commercial licence; Starter at $6 a month adds one. fal charges per character.
- •Free: 10,000 credits a month, no commercial licence listed.
- •Starter $6 a month: 30,000 credits and a commercial licence.
- •fal: $0.04 per thousand characters, no free tier.
ElevenLabs pricing, read
Where to use Eleven v4 Turbo
- ElevenLabs · Free plan and paid tiers; ElevenAgents for conversational use.
- ElevenLabs API · Streaming over WebSocket, where the advertised latency is measured.
- fal · $0.04 per thousand characters, returns a finished file.
- Voyager · Runs inside Voyager, the agent for creative work, on Voyager credits on every paid plan (about 3.6 credits a minute of speech), or with your own fal key.
Voyager
Run Eleven v4 Turbo inside Voyager
Voyager is an agent for creative work that runs every model in this reference, Eleven v4 Turbo included. Brief it, and it picks the model, makes the thing and hands back something you can edit. Create an account to get started.
What we noticed running Eleven v4 Turbo
- •Returned in four to seven seconds per reading, against four to eight for Eleven v4: on whole files from fal the Turbo saving is in price, not wait.
- •22.3 seconds, the same length to the hundredth as Eleven v4's reading, though the files differ. See the names, dates and prices reading.
- •The self-introduction came back 14.03 seconds long, the same length as its ad read to the hundredth.
Eleven v4 Turbo limits and API parameters
| Limit | Value |
|---|---|
| Script length on fal | 5,000 characters a request. fal API reference, read |
| Advertised latency | About 100 ms median inference and 150 ms median to first speech over WebSocket streaming, network latency removed; ElevenLabs' own figures. |
| Voices on fal | Stock voices only; no cloning through fal. fal API reference, read |
›Every setting, for developers
fal (elevenlabs/tts/eleven-v4-turbo)
fal API reference, read
| Parameter | Type | Values | Default |
|---|---|---|---|
| text | string | The script, up to 5,000 characters, with optional [audio tags] and IPA in forward slashes Required. | |
| voice | string | Aria, Roger, Sarah, Laura, Charlie, George, Callum, River, Liam, Charlotte, Alice, Matilda, Will, Jessica, Eric, Chris, Brian, Daniel, Lily, Bill fal's listed default is not among its listed examples. | Rachel |
| stability | number | 0 to 1 Lower is more expressive, higher more consistent. | 0.5 |
| similarity_boost | number | 0 to 1 | 0.75 |
| language_code | string | ISO 639-1 code | |
| apply_text_normalization | enum | auto, on, off Spells out numbers and dates before reading. | auto |
| seed | integer | 0 to 4294967295 Best effort; identical output is not guaranteed. | |
| timestamps | boolean | true, false | false |
| output_format | enum | mp3, PCM, Opus, mu-law and A-law at several rates | mp3_44100_128 |
What Eleven v4 Turbo will not say
ElevenLabs' terms put all use under its Prohibited Use Policy, and a voice model must be your own voice or one you are authorised to share. The free plan's feature list carries no commercial licence; Starter at $6 a month does.
Alternatives to Eleven v4 Turbo
Frequently asked questions
What is Eleven v4 Turbo?
The low-latency variant of ElevenLabs' Eleven v4, released on 28 September 2026 for voice agents and other real-time use, in more than 90 languages.
How much does Eleven v4 Turbo cost?
$0.04 per thousand characters on fal and on ElevenLabs' API, read on 22 October 2026, half of Eleven v4.
Is Turbo faster than Eleven v4?
For streaming, ElevenLabs says so, at about 150 ms to first speech. For whole files from fal, barely: our readings took four to seven seconds against four to eight.
How this page is made
Prices, limits and settings are read from the linked vendor and host pages on the dates shown. Every picture was generated by us through fal with the request shown under it, and the original files are kept unedited. A generation that failed is shown as a failure. Model and vendor names are used to identify the products; no vendor imagery appears on this page.
