Vidu Q3 Mix is a specialized variant focused on multi-entity consistency. Instead of animating a single image, it takes 1–4 reference images and fuses them into a single cohesive video where every subject maintains its distinct identity throughout the clip.
Generate video
All API requests must be authenticated using a Bearer token in the Authorization header. Please ensure your API key is active.Authorization: Bearer sk-xxxxxx
These parameters come from the selected model form_schema. Switching models updates this list and the request example.
images*arrayUpload 1-7 reference images; the model generates a subject-consistent video based on the subjects in them. Supports PNG, JPEG, JPG, WebP; pixels ≥128×128, aspect ratio between 1:4 and 4:1, each image ≤50MB.
prompt*stringText prompt describing the video scene and motion, max 2000 characters.
duration?numberVideo duration in seconds; viduq3-mix range 3-16 seconds, default 5.
aspect_ratio?stringVideo aspect ratio; viduq3-mix supports 16:9, 9:16, 1:1 (4:3 and 3:4 are q2-only). Default 16:9.
resolution?stringVideo output resolution; viduq3-mix supports only 720p and 1080p (no 540p), default 720p.
audio?booleanWhether to use audio-video direct output. viduq3-mix supports synchronized audio; enabled by default. When off, outputs a silent video.
seed?numberRandom seed. Enter a fixed integer for reproducible results; leave empty or set to 0 for a random seed.
watermark?booleanWhether to add a watermark (fixed AI-generated content marker) to the video. Disabled by default.
wm_position?stringWatermark position: 1=top-left, 2=top-right, 3=bottom-right, 4=bottom-left. Effective only when watermark is enabled.
wm_url?stringURL of a custom watermark image. Leave empty to use the default watermark. Effective only when watermark is enabled.
task_id?stringParameter description for Task Id
status?stringParameter description for Status
error?objectParameter description for Error
message?stringError Message
type?stringError Type
param?stringRelated Parameters
code?stringError Code