Which AI video models let you set or clone a voice?
Kling 2.6 took a chosen voice in December 2025 and let elements carry a voice from March 2026. Wan 2.6 reference-to-video kept a person's voice in December 2025 and Wan 2.7 clones a timbre, and Grok Imagine 1.5 offers preset voices. Vidu's lip-sync takes text or audio. From 4 source pages, read 2026-09-25; no maker pays to be listed.
| Date | Line | Version | What changed |
|---|---|---|---|
| 28 July 2025 | Vidu | Not named in the note | Lip-sync from text or audio |
| 3 December 2025 | Wan | Wan 2.6 | Wan 2.6 image-to-video handles dialogue between several speakers |
| 16 December 2025 | Kling AI | Kling 2.6 | Kling 2.6 can speak with a chosen voice |
| 16 December 2025 | Wan | Wan 2.6 | Wan 2.6 reference-to-video keeps a person's look and voice, with several characters at once |
| 23 March 2026 | Kling AI | Not named in the note | Elements made from several images can be tied to a voice |
| 3 April 2026 | Wan | Wan 2.7 | Wan 2.7 reference-to-video mixes up to five image or video references and clones a voice timbre |
| 31 July 2026 | Grok Imagine | Grok Imagine 1.5 | Grok Imagine video 1.5 adds reference-to-video with preset voices and native 1080p |
Inclusion rule. Entries whose note bears directly on the question. Order. By date, oldest first; undated rows last.
Once a model speaks, the next question is whose voice it uses. The notes give three answers: pick one from a set, as with Kling's voice list and Grok's preset voices; carry it with a reference, as with Wan 2.6's reference-to-video; or clone it from audio, as with Wan 2.7 in 2026.
Voice notes cluster from December 2025. Wan 2.6's image-to-video note in the same month adds dialogue between several speakers, which needs a way to keep the voices apart.
1In the makers' words
Vidu, 28 July 2025
Input text/audio to precisely match lip movements in video
Vidu, API update notice
- July 28Lip-sync from text or audio2025
Wan, 3 December 2025
supports stable multi‑speaker dialogue with more natural and realistic vocal timbres
Alibaba Cloud Model Studio, newly released models
- wan2.6-i2v-usWan 2.6 image-to-video handles dialogue between several speakers
Kling AI, 16 December 2025
By using prompt and voice_list, a specified voice can be used to generate videos
Kling AI, API updates
- 12/16/2025Kling 2.6 can speak with a chosen voice
Wan, 16 December 2025
supports using a specified person or any object as a reference, precisely maintaining consistency of appearance and voice, and allows multi‑character reference for joint performances.
Alibaba Cloud Model Studio, newly released models
- wan2.6-r2vWan 2.6 reference-to-video keeps a person's look and voice, with several characters at once
Kling AI, 23 March 2026
Multi-image elements also support binding timbres
Kling AI, API updates
- 03/23/2026Elements made from several images can be tied to a voice
Wan, 3 April 2026
Supports hybrid referencing of up to 5 mixed image/video inputs and audio timbre cloning.
Alibaba Cloud Model Studio, newly released models
- wan2.7-r2vWan 2.7 reference-to-video mixes up to five image or video references and clones a voice timbre
Grok Imagine, 31 July 2026
grok-imagine-video-1.5 now supports text-to-video, image-to-video, and reference-to-video (including optional preset voices), with native 1080p for T2V and I2V.
xAI, API release notes
- July 31Grok Imagine video 1.5 adds reference-to-video with preset voices and native 1080p
Sound, voice and lip-sync · When did AI video models add vertical video? · Which AI video models copy a performance from a video?
2Notes read
- Vidu, API update notice, read 2026-09-25
- Alibaba Cloud Model Studio, newly released models, read 2026-09-25
- Kling AI, API updates, read 2026-09-25
- xAI, API release notes, read 2026-09-25
3Where new versions end up
Release notes such as “Which AI video models let you set or clone a voice?” matter once a pipeline runs the models; SceneMixer's guides describe how it does:
SceneMixer writes dialogue into the script analysis, assigns each line to a shot with a speaker and a short delivery note, and the video model performs it — picture and audio generated together, so the mouth forms that sentence in the language you chose.
SceneMixer, AI Dialogue Generator (read 2026-10-13)
Video is 16:9 landscape by default, with 9:16 vertical available; you set it once per project. Segments are typically 4 to 15 seconds, and some model tiers go to 30.
SceneMixer, AI Short Drama Generator — “What length and aspect ratio does it output?” (read 2026-10-13)