Updated
AlibabaComing soon
Wan 4.0
Alibaba's next video model is due in November 2026.
What is Wan 4.0?
Alibaba has announced its next-generation video model for November 2026, but it has not named it, so Wan 4.0 is a placeholder and the release could be called Wan 3.5 or something else entirely.
At its Apsara Conference on September 22, 2026, Alibaba said the model is still in training and will be longer, more controllable, more complete and smarter. It described director-level creative thinking that moves from single shots to complete narratives, with consistent characters, scenes and shots over longer durations. Wan 3.0, released in August 2026, is the newest Wan model and is live on Flashloop.
- Provider
- Alibaba
- Quality
- To be confirmed
- Aspect ratios
- To be confirmed
- Audio
- To be confirmed
What Wan 4.0 could bring
Longer videos that stay consistent, which Alibaba named as a goal at the Apsara Conference
Precise control over characters, scenes and shots, also stated by Alibaba
Director-level storytelling that moves from single shots to complete narratives, in Alibaba's description
Native audio and mixed references may carry over from Wan 3.0, though Alibaba has not said so
Prompting tips until Wan 4.0 lands
- On Wan 3.0 today, write the clip as a single continuous shot and let the action unfold inside it.
- Say what you want to hear: dialogue in quotes, plus footsteps, ambience or room tone.
- Give each reference one job: an image for the character, a clip for the camera move, an audio file for the tone.
Where Wan 4.0 could fit
Multi-shot stories with recurring characters
Longer narrative clips with planned camera work
Ads and explainers that need shot-by-shot control
How Wan 4.0 compares
| Model | Length | Top resolution | Aspect ratios | Native audio |
|---|---|---|---|---|
| Wan 4.0Coming soon | No | To be confirmed | To be confirmed | To be confirmed |
| Veo 3.1 | 8 seconds | 1080p | 16:9 | Yes |
| Kling 3.0 | 15 seconds | Pro (1080p) | 16:9, 9:16, 1:1 | No |
| Kling 3.0 Turbo | 15 seconds | 1080p | 16:9, 9:16, 1:1 | No |
| Kling 4.0 | 30 seconds | 1080p | 16:9, 9:16, 1:1 +1 | Yes |
Good to know before you generate
- The name is unconfirmed
- No exact date, specs or pricing; Alibaba has only said November
- Still in training when Alibaba announced it in September 2026
Until Wan 4.0 lands on Flashloop
Create with Wan 3.0 todayQuestions about Wan 4.0
Has Alibaba announced Wan 4.0?
Alibaba has announced a next-generation video model for November 2026, but not the name Wan 4.0. Its Apsara Conference update on September 22, 2026, presented the model right after Wan 3.0 and called it only the next-generation video model, without saying whether it will carry the Wan name.
When will Alibaba's next video model come out?
Alibaba said November 2026 and has given no exact day. The model was still in training at the time of the announcement.
What did Alibaba say the new model will do?
It said the model will be longer, more controllable, more complete and smarter. Characters, scenes and shots should stay consistent over longer durations, and the model should think like a director, moving from single shots to understanding a complete narrative. Alibaba gave no resolution, length or audio specs.
Is Iron Goddess on Arena Alibaba's new model?
Nobody knows. An anonymous video model called iron_goddess appeared in Arena video battles around October 18 or 19, 2026. Alibaba is the most common guess, but that is speculation, and no lab has claimed it.
Will it be called Wan 3.5 or Wan 4.0?
Alibaba has not said. Wan versions so far have run 2.1, 2.2, 2.5, 2.6, 2.7 and 3.0, so a 3.5 would match the earlier jump to 2.5, but a 4.0 or a new name is just as possible.
Which Wan models can I use on Flashloop now?
Wan 3.0, Wan 3.0 Prime and Wan 2.6 are live on Flashloop. Wan 3.0 renders 2 to 30 seconds at up to 1080p with native audio.
More video models on Flashloop
Veo 3.1Google DeepMind's Veo 3.1: 8-second videos with dialogue, sound effects and ambience generated in one pass.
Kling 3.0Kuaishou's Kling 3.0: multi-shot video up to 15 seconds, with consistent characters across cuts.
- Kling 3.0 Turbo
Kuaishou's faster, cheaper Kling 3.0: text and image to video at 720p or 1080p, up to 15 seconds.
Kling 4.0Kuaishou's Kling 4.0: clips up to 30 seconds in one pass, with reference control and sound on every clip.