Wan2.6 series (T2V, I2V, R2V, T2I, image)
Wan2.6 adds reference-to-video that keeps a person's look and voice, multi-shot storytelling and clips up to 15 seconds with synchronized audio.
Alibaba pitches role-play video, where users appear as themselves, in their own voice, across several shots. It also offers interleaved text-image output and lengthy bilingual prompts. Available via Model Studio, wan.video and the Qwen app. No open weights.
- Date
- Tuesday, 16 December 2025
- Lab
- Alibaba (Qwen)
- Kind
- model
- Access
- closed API
Figures
| Measure | Value | Measured by |
|---|---|---|
| Max clip length | 15 s Stated in Alibaba's launch release; resolution not specified there | company |
Alibaba Cloud press room dated 2025-12-16; Model Studio lists wan2.6 t2i and image on 2025-12-15, r2v 2025-12-16, i2v-flash 2026-01-15 and r2v-flash 2026-01-29. 'First in China' framing is Alibaba's marketing claim.
Sources
- www.alibabacloud.com/en/press-room/alibaba-unveils-wan2-6-series-enabling-everyone
- www.alibabacloud.com/help/en/model-studio/newly-released-models
This record was checked against its sources on 6 October 2026. How we check