Eleven v4 voice clone: 10 seconds or 1-2 minutes? Pages differ
ElevenLabs' launch post says an instant clone needs 10 seconds of audio; its docs page says one to two minutes. What Sume's Voices clone asks for instead.
How long a sample does an Eleven v4 instant voice clone need? The ElevenLabs launch post says 10 seconds of audio; the Eleven v4 docs page says "generally one to two minutes". Both were read on 2026-10-03. Read the launch post as the minimum that works and the docs page as what to give it for a result you will publish.
Sume's Voices clone asks for a file rather than a duration: wav, mp3, m4a, mp4, webm or ogg, up to 20 MB, plus a name, a gender and a language tag.
Side by side
The launch post also says Professional Voice Clones are supported again in v4 and that instant clones now capture voices with high fidelity from 10 seconds. The product page adds that all clones require verified owner consent. None of the ElevenLabs pages read give a maximum sample length.
| Item | ElevenLabs Eleven v4 | Sume Voices clone |
|---|---|---|
| Sample length | 10 seconds (launch post); one to two minutes (docs page) | No minimum length stated in the product code |
| File types | Not stated on the pages read | wav, mp3, m4a, mp4, webm, ogg |
| Size limit | Not stated on the pages read | 20 MB per source file |
| Required fields | Verified owner consent | Name, gender (male, female, nonbinary), language tag |
| Where the result is used | Eleven v4 and v4 Turbo | As voice.id on a TTS request |
How to choose a sample
Use one speaker only, no music underneath, and the same room and microphone you will use later. A clean 30 to 60 seconds beats a noisy two minutes. Because Sume accepts mp4 and webm, a recorded screen-share with a clear voice track can work, but strip any background music first; audio detach can return a wav of just the track.
Clone only your own voice or one you have written permission to use. Sume's voice id then goes into the voice.id field of tts_create, described in MCP tools and gates.
Sources
Related posts
More in Comparisons
- ElevenLabs v4 Turbo vs Flash vs v3: price per 1,000 characters
ElevenLabs API list per 1K characters: v4 Turbo $0.011 until Oct 12 (then $0.04), Flash/Turbo $0.04, v3 Multilingual $0.08, v4 $0.022 (then $0.08).
- Face and body swap video AI: Recast vs Sume's avatar face swap
Sume has two ways to put someone else in a video: H3 Max Recast swaps the person from a photo, Beta Face Swap applies a ready avatar's face. Which to use.
- Face swap vs H3 Max Recast: which swaps the person in a video?
Sume offers two ways to put a different person in a video: Avatar Face Swap (Beta) and H3 Max Recast. Inputs, length limits, audio and price side by side.
- fal list vs Sume price for Seedance 2.5: what 100 clips cost extra
For 100 five-second 720p Seedance 2.5 clips, fal list is about $231 and Sume is about $289: $57.78 more, the 25 percent house margin. What it covers.
Written by Sume