Eleven v4: 90+ languages or 99? ElevenLabs' own pages differ
ElevenLabs' launch post says more than 90 languages for Eleven v4; its docs page lists 99. How to plan around the gap, and what Sume lists instead.

How many languages does Eleven v4 support? Read on 2026-10-03, the ElevenLabs launch post and product page say "more than 90" and "90+" languages, while the Eleven v4 docs page lists 99. The pages do not contradict each other so much as count differently, and none of them promises the same quality in every language: the docs page says some languages are handled more fluently than others.
For a buying decision, treat 90+ as the safe number and test your own language before you commit. Sume lists fewer languages for its own voices, 16 tags in the Voices library, so the honest comparison is about the languages you actually ship.
What each page says
The launch post is dated September 28, 2026 and was updated October 1, 2026. It says both Eleven v4 and Eleven v4 Turbo support more than 90 languages. The docs page names eleven_v4 and eleven_v4_turbo and lists 99 languages, from Afrikaans to Wolof.
| Page | Count stated | Notes |
|---|---|---|
| Launch post | More than 90 | Both v4 and v4 Turbo |
| Product page (elevenlabs.io/v4) | 90+ | Same wording |
| Docs page | 99 listed | Adds that some languages are more fluent than others |
| Sume Voices library | 16 tags | en, ko, ja, zh, es, fr, de, pt, it, hi, nl, pl, ru, sv, tr, tl |
What to do with the gap
Pick the three languages that carry most of your views and run the same 15-second script in each. If your language is on both ElevenLabs lists and on Sume's 16, compare the two files by ear; if it is only on ElevenLabs' list, the choice is already made.
On Sume, pass language on every non-English request. The tts_create schema says the provider assumes English when the field is omitted, and Sume only infers Korean or Japanese from a Hangul- or kana-only transcript as a fallback. See MCP tools and gates for the full payload.
Sources
Related posts
More in Comparisons
- Eleven v4 drops the native accent; Sume keeps a language tag per voice
ElevenLabs' docs say v4 gives fluent target-language speech, not a preserved accent. Sume tags each voice with a language and checks for mismatch.
- 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.
- 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.
Written by Sume