sora3-vip5 系列视频模型生成视频。API_BASE_URL 和 API Key 由服务商单独提供。API_BASE_URLsora3-vip5-480p | 480p | 854x480 | 480x854 | 请求参数决定,4-15 秒 |sora3-vip5-720p | 720p | 1280x720 | 720x1280 | 请求参数决定,4-15 秒 |sora3-vip5-1080p | 1080p | 1920x1080 | 1080x1920 | 请求参数决定,4-15 秒 |sora3-vip5-fast-480p | 480p | 854x480 | 480x854 | 请求参数决定,4-15 秒 |sora3-vip5-fast-720p | 720p | 1280x720 | 720x1280 | 请求参数决定,4-15 秒 |sora3-vip5-fast-1080p | 1080p | 1920x1080 | 1080x1920 | 请求参数决定,4-15 秒 |sora3-vip5-* 是标准版,质量优先。sora3-vip5-fast-* 是快速版,速度优先。size 切换 480p、720p、1080p。orientation 或 aspect_ratio 明确传递。1. POST /v1/videos 提交视频任务,获取任务 id
2. GET /v1/videos/{id} 轮询任务状态,直到 completed 或 failed
3. 读取 url/video_url 下载生成完成的视频| 参数 | 类型 | 必填 | 说明 |
|---|---|---|---|
model | string | 是 | 模型 ID,只能传本文档列出的 6 个模型之一 |
prompt | string | 是 | 视频提示词 |
duration | int | 否 | 视频时长,支持 4-15 秒;不传默认 4 秒 |
seconds | int | 否 | duration 的兼容字段;两者建议只传一个 |
orientation | string | 否 | landscape 横屏,portrait 竖屏 |
aspect_ratio | string | 否 | 推荐 16:9 或 9:16 |
size | string | 否 | 可传 1280x720、720x1280 等,用于推断横竖屏;不用于切换分辨率档位 |
image_refs | string[] | 否 | 图片参考 URL,最多 9 张 |
audio_refs | string[] | 否 | 音频参考 URL,最多 3 条 |
video_refs | string[] | 否 | 视频参考 URL,最多 3 条 |
generate_audio | boolean | 否 | 是否生成音频,默认 true |
negative_prompt | string | 否 | 负向提示词 |
image_refs 也兼容 imageRefs。audio_refs 也兼容 audioRefs。video_refs 也兼容 videoRefs。| 场景 | 限制 |
|---|---|
| 只 传图片参考 | image_refs 最多 9 张 |
| 传视频参考 | video_refs 最多 3 条,并且至少搭配 1 张 image_refs |
| 传音频参考 | audio_refs 最多 3 条,并且至少搭配 1 张 image_refs |
| 音频 + 图片 | 使用 audio_refs 时,image_refs 最多 2 张 |
| 音频 + 视频 | 使用 audio_refs 时,audio_refs + video_refs 合计最多 3 条 |
| 单个素材大小 | 建议不超过 60MB |
data: URL,但不建议大文件使用 data: 形式。prompt 中可以用 @Image1、@Video1、@Audio1 这类占位符描述素材用途。编号从 1 开始,顺序与数组顺序一致。| 占位符 | 对应字段 | 示例 |
|---|---|---|
@Image1 | image_refs[0] | @Image1 是主角照片 |
@Image2 | image_refs[1] | @Image2 是场景参考 |
@Video1 | video_refs[0] | 参考 @Video1 的动作节奏 |
@Audio1 | audio_refs[0] | 角色使用 @Audio1 的声音 |
{
"model": "sora3-vip5-720p",
"prompt": "一只金毛犬在清晨的海边奔跑,电影感镜头,柔和阳光,真实自然的运动轨迹",
"duration": 8,
"orientation": "landscape",
"generate_audio": true
}{
"model": "sora3-vip5-fast-480p",
"prompt": "竖屏短视频,一名年轻人在城市天台转身看向镜头,背景是夜晚霓虹灯",
"duration": 6,
"aspect_ratio": "9:16"
}{
"model": "sora3-vip5-1080p",
"prompt": "@Image1 是主角形象。让主角在雨夜街道缓慢向前走,镜头从侧后方跟拍,电影质感,高级灰色调",
"duration": 12,
"aspect_ratio": "16:9",
"image_refs": [
"https://example.invalid/assets/character.png"
]
}{
"model": "sora3-vip5-fast-720p",
"prompt": "@Image1 是角色,@Image2 是服装,@Image3 是室内场景。生成角色穿着参考服装走进室内空间的连续镜头,动作自然,镜头平稳推进",
"duration": 15,
"orientation": "landscape",
"image_refs": [
"https://example.invalid/assets/person.png",
"https://example.invalid/assets/outfit.png",
"https://example.invalid/assets/room.png"
]
}{
"model": "sora3-vip5-720p",
"prompt": "@Image1 是主角外观,参考 @Video1 的动作节奏,让主角完成同样的转身和挥手动作",
"duration": 10,
"aspect_ratio": "16:9",
"image_refs": [
"https://example.invalid/assets/person.png"
],
"video_refs": [
"https://example.invalid/assets/motion.mp4"
]
}audio_refs 和 video_refs 合计最多 3 条,并且 image_refs 最多 2 张。{
"model": "sora3-vip5-fast-1080p",
"prompt": "@Image1 是说话人,使用 @Audio1 的声音情绪,参考 @Video1 的镜头运动。生成说话人面对镜头说欢迎语的视频",
"duration": 8,
"orientation": "portrait",
"image_refs": [
"https://example.invalid/assets/speaker.png"
],
"audio_refs": [
"https://example.invalid/assets/voice.wav"
],
"video_refs": [
"https://example.invalid/assets/camera_motion.mp4"
]
}queued 或 processing。{
"id": "video_example_task_id",
"object": "video.task",
"created": 1780000000,
"created_at": 1780000000,
"updated_at": 1780000000,
"model": "sora3-vip5-720p",
"status": "queued",
"progress": 0,
"duration": 8,
"seconds": 8,
"usage": {
"seconds": 8,
"video_count": 1
}
}| 字段 | 说明 |
|---|---|
id | 视频任务 ID,用于查询结果 |
object | 固定为 video.task |
model | 请求使用的公开模型名 |
status | queued、processing、completed、failed |
progress | 进度,整数 0-100 |
duration / seconds | 请求的视频秒数 |
usage.seconds | 本次任务秒数 |
usage.video_count | 本次任务视频数量,固定为 1 |
{
"id": "video_example_task_id",
"object": "video.task",
"created": 1780000000,
"created_at": 1780000000,
"updated_at": 1780000012,
"model": "sora3-vip5-720p",
"status": "processing",
"progress": 35,
"duration": 8,
"seconds": 8,
"usage": {
"seconds": 8,
"video_count": 1
}
}{
"id": "video_example_task_id",
"object": "video.task",
"created": 1780000000,
"created_at": 1780000000,
"updated_at": 1780000180,
"completed_at": 1780000180,
"model": "sora3-vip5-720p",
"status": "completed",
"progress": 100,
"duration": 8,
"seconds": 8,
"usage": {
"seconds": 8,
"video_count": 1
},
"video_url": "https://DOWNLOAD_BASE/generated/example.mp4",
"video": {
"url": "https://DOWNLOAD_BASE/generated/example.mp4"
},
"url": "https://DOWNLOAD_BASE/generated/example.mp4",
"output": {
"url": "https://DOWNLOAD_BASE/generated/example.mp4"
},
"result": {
"url": "https://DOWNLOAD_BASE/generated/example.mp4"
}
}video_urlvideo.urlurloutput.urlresult.url{
"id": "video_example_task_id",
"object": "video.task",
"model": "sora3-vip5-720p",
"status": "failed",
"progress": 0,
"duration": 8,
"seconds": 8,
"usage": {
"seconds": 8,
"video_count": 1
},
"error": {
"message": "生成失败原因"
}
}{
"model": "sora3-vip5-fast-720p",
"prompt": "生成一段横屏城市航拍视频,傍晚金色光线,电影感",
"duration": 8,
"orientation": "landscape",
"stream": false
}{
"id": "chatcmpl_example_id",
"object": "chat.completion",
"created": 1780000180,
"model": "sora3-vip5-fast-720p",
"url": "https://DOWNLOAD_BASE/generated/example.mp4",
"video": {
"url": "https://DOWNLOAD_BASE/generated/example.mp4"
},
"output": "https://DOWNLOAD_BASE/generated/example.mp4",
"result": {
"url": "https://DOWNLOAD_BASE/generated/example.mp4"
},
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "url = https://DOWNLOAD_BASE/generated/example.mp4\n\n```html\n<video src='https://DOWNLOAD_BASE/generated/example.mp4' controls></video>\n```"
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 0,
"completion_tokens": 0,
"total_tokens": 0
}
}stream: true,服务端会持续返回 SSE 心跳,避免客户端误判连接断开。{
"model": "sora3-vip5-1080p",
"prompt": "@Image1 是主角,生成主角在舞台上缓慢走向镜头的视频,聚光灯,电影感",
"duration": 12,
"aspect_ratio": "16:9",
"stream": true,
"image_refs": [
"https://example.invalid/assets/character.png"
]
}data: {"id":"chatcmpl_example_id","object":"chat.completion.chunk","created":1780000000,"model":"sora3-vip5-1080p","choices":[{"index":0,"delta":{"content":""},"finish_reason":null}]}
data: {"id":"chatcmpl_example_id","object":"chat.completion.chunk","created":1780000000,"model":"sora3-vip5-1080p","choices":[{"index":0,"delta":{"content":"url = https://DOWNLOAD_BASE/generated/example.mp4\n\n```html\n<video src='https://DOWNLOAD_BASE/generated/example.mp4' controls></video>\n```"},"finish_reason":null}]}
data: [DONE]messages:{
"model": "sora3-vip5-720p",
"duration": 8,
"stream": true,
"messages": [
{
"role": "user",
"content": [
{
"type": "text",
"text": "参考这张图生成角色转身的视频,镜头从中景推到近景"
},
{
"type": "image_url",
"image_url": {
"url": "https://example.invalid/assets/reference.png"
}
}
]
}
],
"orientation": "portrait"
}audio_refs、video_refs 字段。{
"object": "list",
"data": [
{
"id": "sora3-vip5-480p",
"object": "model",
"owned_by": "service",
"description": "Seedance 2.0 video model (480p, duration via request)"
},
{
"id": "sora3-vip5-fast-480p",
"object": "model",
"owned_by": "service",
"description": "Seedance 2.0 Fast video model (480p, duration via request)"
}
]
}{
"error": {
"message": "prompt is required",
"type": "invalid_request_error"
}
}{
"error": {
"message": "Invalid video model. Use /v1/models to get supported video models",
"type": "invalid_request_error"
}
}{
"error": {
"message": "reference images cannot exceed 9",
"type": "invalid_request_error"
}
}| 状态码 | 说明 |
|---|---|
200 | 请求成功 |
400 | 参数错误,如模型名错误、时长越界、参考素材数量超限 |
401 | API Key 无效或缺失 |
404 | 查询的任务不存在 |
429 | 额度不足或请求过多 |
503 | 服务繁忙、队列已满、排队等待超时或上游临时不可用 |
duration 只支持 4-15 秒,建议传整数。size 只建议用于表达横竖屏,不用于切换清晰度档位。/v1/videos 异步任务接口;如果客户网关容易超时,避免使用长时间非流式同步请求。/v1/chat/completions 时,长任务建议传 stream: true。