Nano Banana

Wan 2.7 — Controllable AI Video Storytelling

Wan2.7-Video, released by Alibaba in April 2026, is built for controllable video storytelling. It reached #1 on the Video-to-Video Arena and entered the Image-to-Video Arena at #5. Here you can use it for text-to-video and reference-to-video.

  • Alibaba Wan2.7-Video
  • Reference-to-video with 1–5 refs
  • Audio and dialogue
  • 2–15 s, 720p / 1080p
Open the full workspace

More models

Why Wan 2.7 Is a Great AI Video Generator

Wan 2.7 is about control: keep a subject and voice from references, direct the shot, and pick the exact length. It suits branded characters, product stories and any scene you need to steer precisely.

  • Built for control

    Alibaba describes Wan2.7-Video as a model for controllable video storytelling, from single clips to full narratives.

  • Arena results

    #1 on the Video-to-Video Arena (1337 Elo) and #5 in Image-to-Video with 1,434 points.

  • Subject references

    Combine up to 5 reference images or clips to keep people and products consistent.

  • Flexible length

    Anywhere from 2 to 15 seconds, so short loops cost less.

What Wan 2.7 can do

  • Text to video

    Describe the scene, action and sound in plain language.

  • Reference to video

    Add images or clips of a subject and place it in a new scene.

  • Sound and voices

    Generates audio with the video, including dialogue.

  • 2–15-second clips

    Pick a length from 2 to 15 seconds, with sound generated in the same pass.

  • 720p and 1080p output

    Choose 720p and 1080p in 5 aspect ratios: 16:9, 9:16, 1:1, 4:3 and 3:4.

  • Reference inputs

    Besides a text prompt you can add up to 5 reference files.

Use cases

Starting points you can copy into the generator above.

Use caseGood forExample prompt
Brand charactersKeep a mascot consistent across clips.

The mascot from the references hands a coffee cup to the camera in a cozy cafe and says "Your usual?" Warm cafe sounds

Product storiesShow a product in use.

The backpack from the reference images on a hiker crossing a rope bridge in a misty forest, creaking rope and birdsong

Short loopsTwo- to five-second loops for web and apps.

A steaming bowl of ramen on a wooden counter, chopsticks lift noodles, soft steam, quiet restaurant hum

Scene draftsBlock out a scene before a shoot.

Two friends talk on a rooftop at sunset, camera slowly circles them, city noise below

Wan 2.7 in 3 steps

  1. 1

    Describe the shot or add a frame

    Write what happens, how the camera moves and how it should sound. Add a start image or references to lock the look.

  2. 2

    Pick length, resolution and ratio

    Set duration, resolution and aspect ratio. The credit cost is shown before you submit.

  3. 3

    Generate and keep your results

    Videos appear in the generator and are saved to My works. Anything that fails is refunded automatically.

X Posts About Wan 2.7 — Controllable AI Video Storytelling

YouTube Videos About Wan 2.7 — Controllable AI Video Storytelling

Frequently asked questions

Text-to-video and reference-to-video (1–5 reference images or videos). Both generate audio.

Ready to Create with Wan 2.7?

1080p clips from text or up to 5 references, with sound, at a moderate credit cost.

Explore All Models
  • Clips up to 15s
  • Native audio
  • Failed outputs refunded