Back to Models

Wan 3.0

By Alibaba

Wan 3.0 in one place: text to video, image to video and reference to video. Pick a mode, then generate with Classic or Agent.

wan/3-0-video-reference-to-video1–39 credits per run

Reference images0/10

Reference videos0/5

Reference audio0/5

Refer to uploads in your prompt as Image1, Video1, Audio1.
Free generation

Wan 3.0 modes

  • Wan 3.0 Text to Video

    Generate a video clip from a text prompt.

    wan/3-0-video-text-to-video

    Use Text to Video
  • Wan 3.0 Image to Video

    Animate an uploaded image into a video clip that follows your prompt.

    wan/3-0-video-image-to-video

    Use Image to Video
  • Wan 3.0 Reference to Video

    Generate a video clip guided by the reference media you upload, referenced by name in the prompt.

    wan/3-0-video-reference-to-video

    Selected above

Reference to Video input parameters

  • Prompt Required

    Describe the video you want to generate.

  • Reference images

    Reference images. Up to 10 files, 20MB each. Cannot be combined with first/last frames.

  • Reference videos Affects credits

    Reference videos. MP4 or MOV, up to 5 files, 100MB each. 1-15s each, 15s total.

  • Reference audio

    Reference audio. MP3 or WAV, up to 5 files, 15MB each. Pair with an image or video.

  • Aspect Ratio Required

    Aspect ratio of the generated video.

    adaptive16:94:31:13:49:16

    Default: adaptive

  • Resolution Required Affects credits

    480P is faster; 1080P is higher quality.

    480P720P1080P

    Default: 1080P

  • Duration Required Affects credits

    Video length in seconds. Use 2–30 without reference video; with reference video, input + output must be ≤ 30s. Pass -1 for smart duration (model chooses length; credits are held at the max allowed output for this call, then excess is refunded after generation).

    -123456789101112131415161718192021222324252627282930

    Default: 5

  • Audio

    Include an audio track in the generated video.

    Default: true

More models from Alibaba

More reference to video models

PoloX AI is an AI creative workspace: chat with agents and work on an infinite canvas to generate images and videos, edit text in images, remove backgrounds, and more.

© 2026 PoloX AI. All rights reserved. polox.ai is a sub-brand of Vision Forge Co., Ltd.

VISIONFORGE CO., LTD — Registered address: 30 N GOULD ST STE R Sheridan, WY 82801.