Alibaba (Qwen) · video model · Released Sep 24, 2025Wan v2.5 Preview
Wan v2.5 Preview is a video model from Alibaba (Qwen), released Sep 24, 2025. It handles text to video and audio input.
- Arena score
- 1,321#22 of 48 · image to video
- Starting price
- —Not published
- Longest clip
- —Not published
- Resolution
- —No audio track
How it rates and what it costs
- Text to video arena
- 1,245 ±9#25 of 48 · 36,378 votes
- Image to video arena
- 1,321 ±10#22 of 48 · 16,907 votes
Arena scores come from blind, side-by-side human votes on the same prompts (higher is better; ± is the 95% range). Boards updated Sep 22, 2026.
The details
- Lab
- Alibaba (Qwen)
- Type
- Video model
- Released
- Sep 24, 2025
- Inputs
- Text prompts, Audio
- Output
- Video
- Can do
- Text to video, Audio input
- Weights
- Proprietary
- Availability
- 1 API host
- Status
- Preview
More video models from Alibaba (Qwen)
About Wan v2.5 Preview
How good is Wan v2.5 Preview?
In blind side-by-side votes, Wan v2.5 Preview scores 1,245 for text to video (#25 of 48) and 1,321 for image to video (#22 of 48). Higher is better; scores within about 20 points of each other are effectively tied.
Who makes Wan v2.5 Preview?
Wan v2.5 Preview is a video model from Alibaba (Qwen), released Sep 24, 2025.
What can Wan v2.5 Preview do?
Wan v2.5 Preview accepts text prompts and audio and produces video. In practice: text to video and audio input.
Is Wan v2.5 Preview open source?
No. Wan v2.5 Preview is proprietary; you use it through an API or partner apps.
Where can I use Wan v2.5 Preview?
It is listed by 1 API host we track.
What are the alternatives to Wan v2.5 Preview?
The closest video models from other labs, by release date, are Sora 2 (OpenAI, Oct 6, 2025), Veo 3.1 (Google, Oct 15, 2025), Kling v2.5 Turbo (Kling AI, Sep 23, 2025) and Seedance v1.0 Pro Fast (ByteDance Seed, Oct 24, 2025).