Kling V3 Omni

Kling V3 Omni 是一款多模态视频模型，支持文生视频、图生视频、首尾帧、多模态参考和视频编辑，并可控制生成音频或保留源视频音频。它适合需要融合图片、视频与声音素材的广告制作、角色和产品一致性内容、镜头改写及带音频的完整短片工作流。你可以在试验场中选择输入素材、时长、清晰度、画面比例和音频选项，在正式接入前比较参考一致性、编辑质量、声音处理、生成速度与成本。具体可用能力与参数范围以当前页面为准。

A multimodal film set with a performer, cinema camera, mirror, and projected ocean imagery

模型代码: kling-v3-omni
供应商: Kling
API 端点: /v1/videos
计费: 0.029411764706
更新时间: 2026-08-13

能力

通过当前公开任务端点生成视频
实时 Schema 提供时支持图片、参考素材、时长或分辨率控制
当前参数包括 prompt, operation, resolution

限制

模型能力与参数范围可能更新，生产环境应锁定并验证使用的参数组合
未出现在实时 Request Schema 中的参数不应假定可用
异步任务需要处理轮询、失败、超时和内容安全状态

典型场景

产品原型和自动化内容工作流
批量媒体生成与程序化创作
需要统一鉴权、计费和调用记录的多模型应用

价格与计费示例

1K/2K image output | 0.029411764706 | output_images | {"operation":["image.generate","image.edit"],"resolution_tier":["std","default","1k","2k"]}
4K image output | 0.058823529412 | output_images | {"operation":["image.generate","image.edit"],"resolution_tier":"4k"}
PRO video editing with audio estimate | 0.205882352941 | input_video_seconds | {"generated_audio":true,"operation":"video.edit","resolution_tier":"pro"}
PRO silent video editing | 0.176470588235 | input_video_seconds | {"generated_audio":false,"operation":"video.edit","resolution_tier":"pro"}
STD video editing with audio estimate | 0.161764705882 | input_video_seconds | {"generated_audio":true,"operation":"video.edit","resolution_tier":"std"}
STD silent video editing | 0.132352941176 | input_video_seconds | {"generated_audio":false,"operation":"video.edit","resolution_tier":"std"}
PRO video with audio without references | 0.147058823529 | output_video_seconds | {"generated_audio":true,"input_videos":0,"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"pro"}
PRO silent video without references | 0.117647058824 | output_video_seconds | {"generated_audio":false,"input_videos":0,"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"pro"}
PRO video with audio and references | 0.205882352941 | output_video_seconds | {"generated_audio":true,"input_videos":{"gte":1},"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"pro"}
PRO silent video with references | 0.176470588235 | output_video_seconds | {"generated_audio":false,"input_videos":{"gte":1},"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"pro"}
STD video with audio without references | 0.117647058824 | output_video_seconds | {"generated_audio":true,"input_videos":0,"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"std"}
STD silent video without references | 0.088235294118 | output_video_seconds | {"generated_audio":false,"input_videos":0,"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"std"}
STD video with audio and references | 0.161764705882 | output_video_seconds | {"generated_audio":true,"input_videos":{"gte":1},"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"std"}
STD silent video with references | 0.132352941176 | output_video_seconds | {"generated_audio":false,"input_videos":{"gte":1},"operation":["video.generate","video.image_to_video","video.reference_to_video"],"resolution_tier":"std"}

请求参数

参数	类型	必填	默认值	可选值
model	string	是
prompt	textarea	是
operation	select	是	text_to_video	text_to_video, image_to_video, first_last_frame, reference_to_video, edit_video
resolution	select	是	std	std, pro
duration	integer	否	5
aspect_ratio	select	否	16:9	16:9, 9:16, 1:1
sound	select	否	off	off, on
image	asset_list	否
last_image	asset_list	否
reference_image_urls	asset_list	否
watermark	boolean	否	false

API 调用示例

cURL

curl -X POST 'https://api.uniall.ai/v1/videos' \
  -H 'Authorization: Bearer $UNIALL_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"aspect_ratio":"16:9","duration":5,"model":"kling-v3-omni","operation":"text_to_video","prompt":"Describe the output you want","resolution":"std","sound":"off","watermark":false}'

Python

import json
import requests

payload = json.loads(r'''{"aspect_ratio":"16:9","duration":5,"model":"kling-v3-omni","operation":"text_to_video","prompt":"Describe the output you want","resolution":"std","sound":"off","watermark":false}''')
response = requests.post(
    "https://api.uniall.ai/v1/videos",
    headers={"Authorization": "Bearer " + UNIALL_API_KEY},
    json=payload,
)
response.raise_for_status()
result = response.json()

JavaScript

const response = await fetch("https://api.uniall.ai/v1/videos", {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.UNIALL_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({"aspect_ratio":"16:9","duration":5,"model":"kling-v3-omni","operation":"text_to_video","prompt":"Describe the output you want","resolution":"std","sound":"off","watermark":false}),
});

if (!response.ok) throw new Error(`HTTP ${response.status}`);
const result = await response.json();

常见问题

kling-v3-omni 的 API 端点是什么？

端点根据当前模型能力和 Request Schema 自动解析，并显示在本页 API 区域。

kling-v3-omni 如何计费？

价格区域直接读取当前结构化计费配置；改价后无需重写本文。

可以使用 Python 调用吗？

可以。本页示例使用当前端点和原始模型代码标识。

是否提供 JavaScript 和 cURL 示例？

提供。三种示例共享同一实时请求结构。

支持哪些分辨率、时长或参考图参数？

支持范围取决于当前 Request Schema，未显示的字段不应假定可用。

生产环境接入时要注意什么？

请验证参数、超时、错误处理和异步视频任务状态，并在模型更新后重新运行小范围测试。