Google
· 发布于 2025-10-16
Veo 3.1
Veo 3.1 is Google’s state-of-the-art video generation model, delivering production-quality video with high visual fidelity and synchronized audio. It creates 1080p videos from text or image prompts with support for dialogue, ambient effects, and background sound. With features including long-scene extension, frame-to-video transitions, vertical video generation, and 4K upscaling, Veo 3.1 enables advanced video creation for storytelling, marketing, and professional creative workflows.
google/veo3.1-preview 官方示例
输入
提示词
这位女士正在接受播客采访,穿着带有标志的蓝紫色上衣,上面写着胜算云的logo,她身处一个中世纪现代风格的工作室,配有白色灯光,她讲述了使用 胜算云并搭配参考图片来制作视频的内容,标志也出现在她身后的黑色背景下有框照片
输出
参数说明
由模型输入 Schema 解析| 参数 | 类型 | 默认值 | 说明 |
|---|---|---|---|
prompt必填 | string | — | 提示词 |
negative_prompt | string | — | 反向提示词 |
aspect_ratio | string | 16:9 | 视频宽高比 |
duration_seconds | integer | 8 | 视频时长(秒) |
resolution | string | 720p | 分辨率 |
generate_audio | boolean | — | 是否生成音频 |
enhance_prompt | boolean | true | 是否使用Gemini优化提示词 |
person_generation | string | allow_adult | 人物生成控制 |
sample_count | integer | 1 | 采样数量 |
seed | integer | -1 | 随机种子 |
image | string | — | 输入图片Google Cloud Storage URI |
reference_images | array | — | 最多3张素材图片或1张风格图片 |
storage_uri | string | — | Google Cloud Storage 存储 URI |
last_frame必填 | string | — | 结束帧图片URL或GCS URI |
接入示例
# 提交任务
curl https://router-integration.test.cogfoundry.ai/api/v1/tasks/generations \
-H "Authorization: Bearer $MODEL_CENTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "google/veo3.1-preview",
"aspect_ratio": "16:9",
"duration_seconds": 8,
"enhance_prompt": true,
"generate_audio": true,
"person_generation": "allow_adult",
"prompt": "这位女士正在接受播客采访,穿着带有标志的蓝紫色上衣,上面写着胜算云的logo,她身处一个中世纪现代风格的工作室,配有白色灯光,她讲述了使用 胜算云并搭配参考图片来制作视频的内容,标志也出现在她身后的黑色背景下有框照片\n",
"reference_images": [
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_1.png",
"reference_type": "ASSET"
},
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_2.png",
"reference_type": "ASSET"
}
],
"resolution": "720p",
"sample_count": 1,
"seed": -1
}'
# 轮询结果(REQUEST_ID 来自上一步返回的 data.request_id)
curl https://router-integration.test.cogfoundry.ai/api/v1/tasks/generations/REQUEST_ID \
-H "Authorization: Bearer $MODEL_CENTER_API_KEY" import requests
BASE = "https://router-integration.test.cogfoundry.ai/api/v1"
KEY = "$MODEL_CENTER_API_KEY"
headers = {"Authorization": f"Bearer {KEY}", "Content-Type": "application/json"}
# 提交任务
resp = requests.post(f"{BASE}/tasks/generations", headers=headers, data='''{
"model": "google/veo3.1-preview",
"aspect_ratio": "16:9",
"duration_seconds": 8,
"enhance_prompt": true,
"generate_audio": true,
"person_generation": "allow_adult",
"prompt": "这位女士正在接受播客采访,穿着带有标志的蓝紫色上衣,上面写着胜算云的logo,她身处一个中世纪现代风格的工作室,配有白色灯光,她讲述了使用 胜算云并搭配参考图片来制作视频的内容,标志也出现在她身后的黑色背景下有框照片\n",
"reference_images": [
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_1.png",
"reference_type": "ASSET"
},
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_2.png",
"reference_type": "ASSET"
}
],
"resolution": "720p",
"sample_count": 1,
"seed": -1
}''')
request_id = resp.json()["data"]["request_id"]
# 轮询结果
result = requests.get(f"{BASE}/tasks/generations/{request_id}", headers=headers)
print(result.json()) const BASE = "https://router-integration.test.cogfoundry.ai/api/v1";
const KEY = "$MODEL_CENTER_API_KEY";
const headers = { Authorization: "Bearer " + KEY, "Content-Type": "application/json" };
// 提交任务
const payload = {
"model": "google/veo3.1-preview",
"aspect_ratio": "16:9",
"duration_seconds": 8,
"enhance_prompt": true,
"generate_audio": true,
"person_generation": "allow_adult",
"prompt": "这位女士正在接受播客采访,穿着带有标志的蓝紫色上衣,上面写着胜算云的logo,她身处一个中世纪现代风格的工作室,配有白色灯光,她讲述了使用 胜算云并搭配参考图片来制作视频的内容,标志也出现在她身后的黑色背景下有框照片\n",
"reference_images": [
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_1.png",
"reference_type": "ASSET"
},
{
"image": "https://oss.shengsuanyun.com/example/modelinfo/201/2025-11-25_13:48:44_input_2.png",
"reference_type": "ASSET"
}
],
"resolution": "720p",
"sample_count": 1,
"seed": -1
};
const resp = await fetch(BASE + "/tasks/generations", {
method: "POST",
headers,
body: JSON.stringify(payload),
});
const { data } = await resp.json();
// 轮询结果
const result = await fetch(BASE + "/tasks/generations/" + data.request_id, { headers });
console.log(await result.json()); 定价
USD · 实时汇率| 定价 | USD |
|---|---|
| 不带音频 | $0.20 |
| 带音频 | $0.40 |
供应商
提供该模型的渠道| 供应商 | 起价 |
|---|---|
| VertexTask | $0.20 |