跳到主要内容
MiniMax检查中

MiniMax H3 / Multimodal Reference to Video

minimax-h3-multimodal-reference-to-video

reference to video输出: 视频输入: 文本 + 图像 + 视频 + audio
正在加载价格…
官方原价 · 未验证

模型能力

Generate video using images, video and audio references.

768p / 2K · multimodal references

输入参数

仅列出此模型接受的字段。必填字段必须提供。

字段类型必填默认值可选值与约束
prompt
string必填—

Text prompt for video generation.

1–20480 chars

imageUrls
string[]可选—

reference images (max 9)

≤ 9 items · HTTPS image URL or media:<uuid>

videoUrls
string[]可选—

reference video

≤ 3 items · HTTPS video URL or media:<uuid>

audioUrls
string[]可选—

reference audio

≤ 3 items · HTTPS audio URL or media:<uuid>

resolution
string必填"768P"

Output resolution.

可选值: "2K", "768P"

duration
string必填"5"

Duration in seconds, 5-15.

可选值: "5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15"

ratio
string可选"adaptive"

Aspect ratio, default adaptive.

可选值: "adaptive", "21:9", "16:9", "4:3", "1:1", "3:4", "9:16"

aigcWatermark
boolean可选—

Whether to add an AIGC watermark to the generated video.

限制与行为

  • 仅支持异步:先提交任务,再轮询结果。
  • prompt: 1–20480 chars
  • imageUrls: ≤ 9 items
  • videoUrls: ≤ 3 items
  • audioUrls: ≤ 3 items
  • 媒体输入接受公开 HTTPS URL,或 POST /api/v1/media 返回的 media:<uuid> 引用。