视频模型 minimax
- 2026.09.23 本站已经支持 视频模型
minimax-h3官方 api 格式 - 支持
文生视频图生视频多模态视频 - 支持
H3-Context-IR
Minimax API
- 支持
文生视频图生视频多模态视频 - 官方
baseurl:https://api.minimax.cn - 我们的
baseurl:https://api.openai-hk.com/minimax - 官方文档
1.创建任务
- 返回
task_id
json
{
"task_id": "2102690106093002752"
}{
"task_id": "2102690106093002752"
}1.1.文生视频
shell
curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "必须遵循以下提示词,不允许改动:\u2028实景质感的中国古代修仙场景与手绘发光动画融合的实验性影像。夜晚古朴的观星阁,月光透过雕花木窗洒下。画面中心是一位绝美仙女(真实人物摄影质感):肤如凝脂,眉若远山,双眸如秋水般清透,眼角点缀着微红的精致花钿;梳着古典飞仙髻,斜插一根温润的白玉簪;身穿轻盈透气的交领广袖流仙裙(层叠的半透明月白色与淡青色丝绸),臂弯挽着随风飘动的轻纱披帛。采用像用手机随手记录一样的强烈手持晃动感、靠近人物时的对焦延迟和边缘暗角。不要处理成精致完美的电影画面,而是像深夜偶然捕捉到仙女与奇妙法术现象嬉戏的真实粗糙记录质感。\n0〜3秒:镜头微微晃动,中近景对准仙女清冷美丽的脸庞。她嘴角带着一抹温柔的浅笑,向镜头前方伸出白皙纤细的真实手掌。在她掌心上方,凭空出现了一颗手绘发光动画风格的“金色小光种”。光种像粉笔画一样带有毛边,顺着她掌心的纹理打转,随后瞬间炸开,变成一只金灿灿的“手绘小灵蝶”,顺着她的指尖飞起。镜头稍有延迟地跟随着灵蝶向上摇摄(Tilt)。\n3〜6秒:金色小灵蝶在仙女的白玉簪旁翩翩起舞,仙女微微侧头,广袖流仙裙的丝绸泛起真实的柔和光泽。当她抬起手用宽大的衣袖轻轻扑打时,小灵蝶顺势落在了她的青色袖口上,瞬间拉长,变成了一片扁平的发光“手绘小竹叶”。竹叶像小舟一样飘起,拖着淡青色的流光轨迹,飞向仙女的肩膀。镜头慢半拍地绕着仙女微微移动追踪。\n6〜10秒:青色手绘竹叶触碰到仙女飘动的轻纱披帛,顺着纱质的纹理滑行,变成了一条手绘的“半透明发光锦鲤”。锦鲤在仙女周身游动,仙女新奇地睁大眼睛,清透的眼眸中倒映着锦鲤的光芒。锦鲤绕到她的脑后,立刻变成了一条灵动的“手绘发光水墨小龙”,顺着仙女的肩膀游走。小龙吐出一团迷你的光雾,向远处的青铜宫灯逃去。镜头带有真实的晕眩感,跟随着小龙向右平移。\n10〜13秒:手绘小龙撞到了青铜宫灯上,周围爆开一圈金绿交织的手绘小八卦阵。仙女好奇地倾身靠近(此时镜头极近距离展现她精致的五官、根根分明的睫毛与肌肤纹理),小龙瞬间变成一个长着草叶帽子的“手绘小妖精”。小妖精调皮地揪了一下仙女的一缕真实发丝,然后慌忙跳入半空,留下一串点状的流光脚印,逃向阁楼的木柱。镜头带有略微的失焦感,追赶小妖精。\n13〜15秒:点状脚印碰到木柱的瞬间,金、绿、青色的手绘线条疯狂蔓延,覆盖了窗户、墙壁和仙女的裙摆。整个观星阁瞬间变成了一幅巨大的“发光星宿山水图”手绘涂鸦。最后,仙女转头看向镜头,用宽大的袖子半掩着面容发出一阵轻笑,一片手绘的粉色发光小桃花瓣飘落,正好贴在她挺拔的鼻尖上。\n手绘动画部分绝对不能是平滑的CG,必须是像1帧1帧重新绘制一样的粗糙抽帧动画(Frame-by-frame animation)。线条粗细、轮廓和上色始终有微妙的沸腾抖动,保留蜡笔、粉笔、彩色铅笔、粗糙毛笔的纹理。以金黄、淡青、翠绿、朱红为主的东方仙侠柔和发光色。动画周围有微弱的光晕,但不能把现实场景照得过亮。\n镜头必须始终在动画之后“被动跟随”。动画先移出画框,手持镜头才迟缓地追上去。绝对禁止使用平稳的稳定器运动。禁止过于精密的3DCG、禁止毛绒玩具质感、禁止过于平滑的矢量线、禁止均匀死板的霓虹灯效果、禁止电影般完美的布光、禁止过度的景深、禁止广告般完美的构图,禁止字幕、Logo和BGM。禁止出现巨大的眼睛、裂开的嘴、獠牙、扑脸动作、突然黑屏和跳跃惊吓。\n仅保留室内的白噪音、轻柔的夜风声、仙女衣袂翻飞的真实丝绸摩擦声、玉饰碰撞的清脆声响、仙女极其细微的呼吸与轻笑声,以及手绘小妖精发出柔和的“叮当”法术音和像小鸟一样的嘟囔声。"
}
],
"resolution": "768p",
"duration": 15,
"ratio": "16:9"
}'curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "必须遵循以下提示词,不允许改动:\u2028实景质感的中国古代修仙场景与手绘发光动画融合的实验性影像。夜晚古朴的观星阁,月光透过雕花木窗洒下。画面中心是一位绝美仙女(真实人物摄影质感):肤如凝脂,眉若远山,双眸如秋水般清透,眼角点缀着微红的精致花钿;梳着古典飞仙髻,斜插一根温润的白玉簪;身穿轻盈透气的交领广袖流仙裙(层叠的半透明月白色与淡青色丝绸),臂弯挽着随风飘动的轻纱披帛。采用像用手机随手记录一样的强烈手持晃动感、靠近人物时的对焦延迟和边缘暗角。不要处理成精致完美的电影画面,而是像深夜偶然捕捉到仙女与奇妙法术现象嬉戏的真实粗糙记录质感。\n0〜3秒:镜头微微晃动,中近景对准仙女清冷美丽的脸庞。她嘴角带着一抹温柔的浅笑,向镜头前方伸出白皙纤细的真实手掌。在她掌心上方,凭空出现了一颗手绘发光动画风格的“金色小光种”。光种像粉笔画一样带有毛边,顺着她掌心的纹理打转,随后瞬间炸开,变成一只金灿灿的“手绘小灵蝶”,顺着她的指尖飞起。镜头稍有延迟地跟随着灵蝶向上摇摄(Tilt)。\n3〜6秒:金色小灵蝶在仙女的白玉簪旁翩翩起舞,仙女微微侧头,广袖流仙裙的丝绸泛起真实的柔和光泽。当她抬起手用宽大的衣袖轻轻扑打时,小灵蝶顺势落在了她的青色袖口上,瞬间拉长,变成了一片扁平的发光“手绘小竹叶”。竹叶像小舟一样飘起,拖着淡青色的流光轨迹,飞向仙女的肩膀。镜头慢半拍地绕着仙女微微移动追踪。\n6〜10秒:青色手绘竹叶触碰到仙女飘动的轻纱披帛,顺着纱质的纹理滑行,变成了一条手绘的“半透明发光锦鲤”。锦鲤在仙女周身游动,仙女新奇地睁大眼睛,清透的眼眸中倒映着锦鲤的光芒。锦鲤绕到她的脑后,立刻变成了一条灵动的“手绘发光水墨小龙”,顺着仙女的肩膀游走。小龙吐出一团迷你的光雾,向远处的青铜宫灯逃去。镜头带有真实的晕眩感,跟随着小龙向右平移。\n10〜13秒:手绘小龙撞到了青铜宫灯上,周围爆开一圈金绿交织的手绘小八卦阵。仙女好奇地倾身靠近(此时镜头极近距离展现她精致的五官、根根分明的睫毛与肌肤纹理),小龙瞬间变成一个长着草叶帽子的“手绘小妖精”。小妖精调皮地揪了一下仙女的一缕真实发丝,然后慌忙跳入半空,留下一串点状的流光脚印,逃向阁楼的木柱。镜头带有略微的失焦感,追赶小妖精。\n13〜15秒:点状脚印碰到木柱的瞬间,金、绿、青色的手绘线条疯狂蔓延,覆盖了窗户、墙壁和仙女的裙摆。整个观星阁瞬间变成了一幅巨大的“发光星宿山水图”手绘涂鸦。最后,仙女转头看向镜头,用宽大的袖子半掩着面容发出一阵轻笑,一片手绘的粉色发光小桃花瓣飘落,正好贴在她挺拔的鼻尖上。\n手绘动画部分绝对不能是平滑的CG,必须是像1帧1帧重新绘制一样的粗糙抽帧动画(Frame-by-frame animation)。线条粗细、轮廓和上色始终有微妙的沸腾抖动,保留蜡笔、粉笔、彩色铅笔、粗糙毛笔的纹理。以金黄、淡青、翠绿、朱红为主的东方仙侠柔和发光色。动画周围有微弱的光晕,但不能把现实场景照得过亮。\n镜头必须始终在动画之后“被动跟随”。动画先移出画框,手持镜头才迟缓地追上去。绝对禁止使用平稳的稳定器运动。禁止过于精密的3DCG、禁止毛绒玩具质感、禁止过于平滑的矢量线、禁止均匀死板的霓虹灯效果、禁止电影般完美的布光、禁止过度的景深、禁止广告般完美的构图,禁止字幕、Logo和BGM。禁止出现巨大的眼睛、裂开的嘴、獠牙、扑脸动作、突然黑屏和跳跃惊吓。\n仅保留室内的白噪音、轻柔的夜风声、仙女衣袂翻飞的真实丝绸摩擦声、玉饰碰撞的清脆声响、仙女极其细微的呼吸与轻笑声,以及手绘小妖精发出柔和的“叮当”法术音和像小鸟一样的嘟囔声。"
}
],
"resolution": "768p",
"duration": 15,
"ratio": "16:9"
}'1.2.图生视频
shell
curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "明亮鲜艳的商业广告片特写,主角是新鲜的新疆吐鲁番葡萄。紧凑的微距镜头下,串串饱满圆润的葡萄挂着晶莹水珠,透着翠绿与深紫的诱人光泽。极慢镜头下,一颗葡萄被轻压爆发挤出浓郁甘甜的果汁,水花与果汁在明亮阳光下飞溅爆发。背景干净透亮,色彩高饱和极具冲击力,光影璀璨。葡萄参考 @Image1"
},
{
"type": "image_url",
"image_url": {
"url": "https://myfile.aibeke.com/res/tutao.webp"
},
"role": "reference_image"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "16:9"
}'curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "明亮鲜艳的商业广告片特写,主角是新鲜的新疆吐鲁番葡萄。紧凑的微距镜头下,串串饱满圆润的葡萄挂着晶莹水珠,透着翠绿与深紫的诱人光泽。极慢镜头下,一颗葡萄被轻压爆发挤出浓郁甘甜的果汁,水花与果汁在明亮阳光下飞溅爆发。背景干净透亮,色彩高饱和极具冲击力,光影璀璨。葡萄参考 @Image1"
},
{
"type": "image_url",
"image_url": {
"url": "https://myfile.aibeke.com/res/tutao.webp"
},
"role": "reference_image"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "16:9"
}'1.3.图生视频
shell
curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "subject_definitions:\n<Subject 1> is the man in <Video 1>, who has curly blonde hair and wears a bright pink suit jacket over a white shirt, holding a black lamb.\n<Audio 1> is the voice timbre reference for <Subject 1>'\''s voice, containing a spoken English line.\n\nsummary:\n[reference generation + audio reference] The target video features <Subject 1> standing in a grassy field at sunset, holding a black lamb while several white lambs graze in the background. He looks up at the sky and speaks a poetic phrase about freedom, with his voice timbre guided by <Audio 1>.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the target video preserves his identity, curly blonde hair, pink suit jacket, white shirt, and the black lamb he holds.\n<Audio 1>: reference - the target audio references the voice timbre and delivery style of <Audio 1> for the character'\''s new dialogue, without copying its original lexical content.\n\ndetailed_description:\nThe target video is in realistic cinematic style.\n\n[Shot 1] The shot opens with a medium view of <Subject 1> standing in a lush green grassy field under a clear blue sky during golden hour. <Subject 1> is a man with curly blonde hair, wearing a bright pink suit jacket over an unbuttoned white shirt. He is holding a small black lamb securely against his chest. In the background, several white lambs are grazing peacefully on the hillside. The warm sunset lighting casts soft shadows across the landscape and highlights the texture of his pink suit. Gentle wind blows through his hair and rustles the tall grass. The camera slowly pushes in toward him. As the scene unfolds, only the ambient sound of the wind and distant nature is audible. <Subject 1> lifts his chin slightly, looking up toward the sky with a serene, contemplative expression. His hands gently adjust their grip on the black lamb. <Subject 1> (S1), speaking with a calm, resonant, and poetic voice matching the timbre of <Audio 1>, says, <d>[English] Follow the wind, live free.</d> He takes a slow breath, his eyes still gazing upward as the breeze continues to blow. <Subject 1> (S1) continues softly, <d>[English] Leave worries behind, enjoy the moment.</d> After he finishes speaking, his lips close into a relaxed smile. He lowers his gaze from the sky, looking down affectionately at the black lamb in his arms, and begins to gently stroke its dark fur with his right hand. The camera maintains its steady push-in as the ambient wind and quiet outdoor atmosphere remain audible through the end of the shot.\n\noverall_soundscape:\nThe soundscape features a tranquil outdoor atmosphere with gentle wind blowing across the grassy field, accompanied by the character'\''s calm, poetic speech.\n\nnon_diegetic_music:\nN/A"
},
{
"type": "video_url",
"video_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/297573323635_00_视频1_YnyRbxEwio_video_20260525_163755_1927e9d3.mp4"
},
"role": "reference_video"
},
{
"type": "audio_url",
"audio_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/f463d523c5ce_01_音频1_RSLcbpzJPo_6月5日(1).mp3"
},
"role": "reference_audio"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "adaptive"
}'curl --request POST \
--url https://api.openai-hk.com/minimax/v2/video_generation \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json' \
--data '{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "subject_definitions:\n<Subject 1> is the man in <Video 1>, who has curly blonde hair and wears a bright pink suit jacket over a white shirt, holding a black lamb.\n<Audio 1> is the voice timbre reference for <Subject 1>'\''s voice, containing a spoken English line.\n\nsummary:\n[reference generation + audio reference] The target video features <Subject 1> standing in a grassy field at sunset, holding a black lamb while several white lambs graze in the background. He looks up at the sky and speaks a poetic phrase about freedom, with his voice timbre guided by <Audio 1>.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the target video preserves his identity, curly blonde hair, pink suit jacket, white shirt, and the black lamb he holds.\n<Audio 1>: reference - the target audio references the voice timbre and delivery style of <Audio 1> for the character'\''s new dialogue, without copying its original lexical content.\n\ndetailed_description:\nThe target video is in realistic cinematic style.\n\n[Shot 1] The shot opens with a medium view of <Subject 1> standing in a lush green grassy field under a clear blue sky during golden hour. <Subject 1> is a man with curly blonde hair, wearing a bright pink suit jacket over an unbuttoned white shirt. He is holding a small black lamb securely against his chest. In the background, several white lambs are grazing peacefully on the hillside. The warm sunset lighting casts soft shadows across the landscape and highlights the texture of his pink suit. Gentle wind blows through his hair and rustles the tall grass. The camera slowly pushes in toward him. As the scene unfolds, only the ambient sound of the wind and distant nature is audible. <Subject 1> lifts his chin slightly, looking up toward the sky with a serene, contemplative expression. His hands gently adjust their grip on the black lamb. <Subject 1> (S1), speaking with a calm, resonant, and poetic voice matching the timbre of <Audio 1>, says, <d>[English] Follow the wind, live free.</d> He takes a slow breath, his eyes still gazing upward as the breeze continues to blow. <Subject 1> (S1) continues softly, <d>[English] Leave worries behind, enjoy the moment.</d> After he finishes speaking, his lips close into a relaxed smile. He lowers his gaze from the sky, looking down affectionately at the black lamb in his arms, and begins to gently stroke its dark fur with his right hand. The camera maintains its steady push-in as the ambient wind and quiet outdoor atmosphere remain audible through the end of the shot.\n\noverall_soundscape:\nThe soundscape features a tranquil outdoor atmosphere with gentle wind blowing across the grassy field, accompanied by the character'\''s calm, poetic speech.\n\nnon_diegetic_music:\nN/A"
},
{
"type": "video_url",
"video_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/297573323635_00_视频1_YnyRbxEwio_video_20260525_163755_1927e9d3.mp4"
},
"role": "reference_video"
},
{
"type": "audio_url",
"audio_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/f463d523c5ce_01_音频1_RSLcbpzJPo_6月5日(1).mp3"
},
"role": "reference_audio"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "adaptive"
}'2. 提交体主要参数说明
| 字段 | 必填 | 说明 |
|---|---|---|
model | 是 | 上表中的模型名 |
content | 是 | 提示词 + 参考素材数组,见下 |
duration | 否 | 时长(秒) |
resolution | 否 | 分辨率,如 768p / 2k |
ratio | 否 | 画幅,如 16:9(别名 aspect_ratio) |
content 数组元素:
type | 取素材的字段 | role | 说明 |
|---|---|---|---|
text | text | — | 提示词。取第一条非空 text |
image_url | image_url.url | reference_image(默认) | 参考图,对应 @Image1、@Image2… |
image_url | image_url.url | first_frame / last_frame | 首帧 / 尾帧图,不参与 @ImageN 编号 |
video_url | video_url.url | reference_video | 参考视频,对应 @Video1、@Video2… |
audio_url | audio_url.url | reference_audio | 参考音频 |
3.获取任务结果
get https://api.openai-hk.com//minimax/v2/query/video_generation/{taskId}
shell
curl --request GET \
--url https://api.openai-hk.com/minimax/v2/query/video_generation/2102690106093002752 \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json'curl --request GET \
--url https://api.openai-hk.com/minimax/v2/query/video_generation/2102690106093002752 \
--header 'Authorization: Bearer hk-your-key' \
--header 'Content-Type: application/json'- 返回
json
{
"task": {
"content": {
"url": "https://suno.day/minimax/2102690106093002752/content?expires=1790760224&signature=dac1dc8e614e5f9eb852f33c06d7d534e11a2fa048e2c5236aaa7446cc94482b"
},
"created_at": 1790155362,
"duration": 15,
"estimated_remaining_seconds": 0,
"id": "2102690106093002752",
"modality": "video",
"model": "MiniMax-H3",
"ratio": "16:9",
"resolution": "768P",
"status": "succeeded",
"task_type": "generation",
"title": "观星阁夜遇仙灵手绘奇缘",
"updated_at": 1790155415,
"usage": {
"input_image_count": 0,
"input_seconds": 0,
"output_seconds": 15,
"total_seconds": 15
}
}
}{
"task": {
"content": {
"url": "https://suno.day/minimax/2102690106093002752/content?expires=1790760224&signature=dac1dc8e614e5f9eb852f33c06d7d534e11a2fa048e2c5236aaa7446cc94482b"
},
"created_at": 1790155362,
"duration": 15,
"estimated_remaining_seconds": 0,
"id": "2102690106093002752",
"modality": "video",
"model": "MiniMax-H3",
"ratio": "16:9",
"resolution": "768P",
"status": "succeeded",
"task_type": "generation",
"title": "观星阁夜遇仙灵手绘奇缘",
"updated_at": 1790155415,
"usage": {
"input_image_count": 0,
"input_seconds": 0,
"output_seconds": 15,
"total_seconds": 15
}
}
}task.content.url返回的是一个mp4文件,可直接下载播放。task.status返回的状态为succeeded、failedqueuedrunningtask.usage返回的是用量信息,用于计费
4.H3-Context-IR 提示词优化
H3-Context-IR 对文本、图像、音频和视频等多模态上下文进行深度理解,分析素材之间以及素材与目标生成结果之间的关系,并进行复杂逻辑推理。系统会将理解结果转换为结构化表达,在尽量保持用户原始意图的前提下丰富语义细节。
计费 1000 积分一次
4.1.创建任务
- 请求体
json
{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "Character speaks: Follow the wind, live free. Leave worries behind, enjoy the moment. Voice timbre follows reference audio 1."
},
{
"type": "video_url",
"video_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/297573323635_00_%E8%A7%86%E9%A2%911_YnyRbxEwio_video_20260525_163755_1927e9d3.mp4"
},
"role": "reference_video"
},
{
"type": "audio_url",
"audio_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/f463d523c5ce_01_%E9%9F%B3%E9%A2%911_RSLcbpzJPo_6%E6%9C%885%E6%97%A5(1).mp3"
},
"role": "reference_audio"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "adaptive"
}{
"model": "MiniMax-H3",
"content": [
{
"type": "text",
"text": "Character speaks: Follow the wind, live free. Leave worries behind, enjoy the moment. Voice timbre follows reference audio 1."
},
{
"type": "video_url",
"video_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/297573323635_00_%E8%A7%86%E9%A2%911_YnyRbxEwio_video_20260525_163755_1927e9d3.mp4"
},
"role": "reference_video"
},
{
"type": "audio_url",
"audio_url": {
"url": "https://cdn.hailuoai.com/prod/hailuo_demo/testsets/h3_promo_eval_ref2va/gallery/sr_v2p26_trio_seed42_20260724/inputs/f463d523c5ce_01_%E9%9F%B3%E9%A2%911_RSLcbpzJPo_6%E6%9C%885%E6%97%A5(1).mp3"
},
"role": "reference_audio"
}
],
"resolution": "768p",
"duration": 5,
"ratio": "adaptive"
}- 返回
task_id
4.2.获取结果
get https://api.openai-hk.com/minimax/v2/query/video_generation/{taskId}
- 返回结果
- 结果
task.task_type:h3_context_ir - 结果
task.content.prompt为优化好的提示词
json
{
"task": {
"id": "2102587448852533248",
"model": "MiniMax-H3",
"title": "",
"status": "succeeded",
"created_at": 1790130887,
"updated_at": 1790130890,
"content": {
"prompt": "subject_definitions:\n<Subject 1> is the young man with curly blonde hair wearing a bright pink suit jacket, matching trousers, and an unbuttoned white shirt in <Video 1>.\n<Video 1> is the source video for the editing baseline.\n<Audio 1> is the voice timbre reference for <Subject 1>'s voice, containing a spoken English dialogue layer.\n\nsummary:\n[video editing + audio reference] The target video is an edited version of <Video 1>. <Subject 1> stands on a grassy hill holding a black lamb while white lambs graze nearby. As he looks out at the landscape, he speaks a poetic line about freedom, using the voice timbre from <Audio 1>.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the young man's identity, curly blonde hair, bright pink suit jacket, matching trousers, and unbuttoned white shirt are completely retained.\n<Video 1> (source-video editing): partially_preserved - the source video's framing, camera push-in, lighting, and background environment are preserved, but <Subject 1> now actively speaks to the viewer instead of remaining silent.\n<Audio 1>: reference - the deep, resonant voice timbre and dramatic delivery style from <Audio 1> are used as the vocal profile for <Subject 1>'s new dialogue, without reusing its original Shakespearean words.\n\ndetailed_description:\nThe target video is in realistic live-action style.\n[Shot 1] The shot begins with <Subject 1>, a young man with curly blonde hair wearing a bright pink suit jacket, matching trousers, and an unbuttoned white shirt, standing prominently in the center foreground on a lush green grassy hill. He holds a small, fluffy black lamb securely against his chest with both hands. In the midground behind him, several white lambs graze peacefully on the sloping terrain under a clear blue sky dotted with faint clouds. Warm, golden sunlight bathes the scene, highlighting the texture of his hair and the fabric of his pink suit. The camera slowly pushes in toward <Subject 1>. Initially, only the gentle ambient sound of wind and nature is audible. As the camera moves closer, <Subject 1> lifts his chin, looking out over the expansive landscape with a calm, liberated expression. The young man (S1) opens his mouth and speaks with a deep, resonant, and slightly gravelly tone, <d>[English] Follow the wind, live free.</d> He maintains his steady gaze into the distance, gently adjusting his grip on the black lamb. Continuing his thought, the young man (S1) adds, <d>[English] Leave worries behind, enjoy the moment.</d> Immediately after his final word, his lips close into a relaxed position, and his jaw ceases speaking motion. His head turns slightly further along the horizon, and his shoulders visibly drop as he takes a slow, visible breath, fully surrendering to the peaceful atmosphere. The quiet natural ambience remains audible through the end of the shot.\n\noverall_soundscape:\nThe soundscape features the gentle ambient sound of outdoor wind and nature, centered around the young man's clear, resonant spoken dialogue.\n\nnon_diegetic_music:\nN/A"
},
"duration": 5,
"usage": {
"total_tokens": 4605,
"prompt_tokens": 3910,
"completion_tokens": 695
},
"ratio": "adaptive",
"task_type": "h3_context_ir",
"modality": "text",
"estimated_remaining_seconds": 0
}
}{
"task": {
"id": "2102587448852533248",
"model": "MiniMax-H3",
"title": "",
"status": "succeeded",
"created_at": 1790130887,
"updated_at": 1790130890,
"content": {
"prompt": "subject_definitions:\n<Subject 1> is the young man with curly blonde hair wearing a bright pink suit jacket, matching trousers, and an unbuttoned white shirt in <Video 1>.\n<Video 1> is the source video for the editing baseline.\n<Audio 1> is the voice timbre reference for <Subject 1>'s voice, containing a spoken English dialogue layer.\n\nsummary:\n[video editing + audio reference] The target video is an edited version of <Video 1>. <Subject 1> stands on a grassy hill holding a black lamb while white lambs graze nearby. As he looks out at the landscape, he speaks a poetic line about freedom, using the voice timbre from <Audio 1>.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the young man's identity, curly blonde hair, bright pink suit jacket, matching trousers, and unbuttoned white shirt are completely retained.\n<Video 1> (source-video editing): partially_preserved - the source video's framing, camera push-in, lighting, and background environment are preserved, but <Subject 1> now actively speaks to the viewer instead of remaining silent.\n<Audio 1>: reference - the deep, resonant voice timbre and dramatic delivery style from <Audio 1> are used as the vocal profile for <Subject 1>'s new dialogue, without reusing its original Shakespearean words.\n\ndetailed_description:\nThe target video is in realistic live-action style.\n[Shot 1] The shot begins with <Subject 1>, a young man with curly blonde hair wearing a bright pink suit jacket, matching trousers, and an unbuttoned white shirt, standing prominently in the center foreground on a lush green grassy hill. He holds a small, fluffy black lamb securely against his chest with both hands. In the midground behind him, several white lambs graze peacefully on the sloping terrain under a clear blue sky dotted with faint clouds. Warm, golden sunlight bathes the scene, highlighting the texture of his hair and the fabric of his pink suit. The camera slowly pushes in toward <Subject 1>. Initially, only the gentle ambient sound of wind and nature is audible. As the camera moves closer, <Subject 1> lifts his chin, looking out over the expansive landscape with a calm, liberated expression. The young man (S1) opens his mouth and speaks with a deep, resonant, and slightly gravelly tone, <d>[English] Follow the wind, live free.</d> He maintains his steady gaze into the distance, gently adjusting his grip on the black lamb. Continuing his thought, the young man (S1) adds, <d>[English] Leave worries behind, enjoy the moment.</d> Immediately after his final word, his lips close into a relaxed position, and his jaw ceases speaking motion. His head turns slightly further along the horizon, and his shoulders visibly drop as he takes a slow, visible breath, fully surrendering to the peaceful atmosphere. The quiet natural ambience remains audible through the end of the shot.\n\noverall_soundscape:\nThe soundscape features the gentle ambient sound of outdoor wind and nature, centered around the young man's clear, resonant spoken dialogue.\n\nnon_diegetic_music:\nN/A"
},
"duration": 5,
"usage": {
"total_tokens": 4605,
"prompt_tokens": 3910,
"completion_tokens": 695
},
"ratio": "adaptive",
"task_type": "h3_context_ir",
"modality": "text",
"estimated_remaining_seconds": 0
}
}计费与模型
- 价格与模型 当前仅支持
MiniMax-H3模型,768p为 2000 积分/s 、2k为 3000 积分/s - 根据
usage.input_seconds涉及输入的参考视频时长 - 根据
usage.output_seconds涉及输出的视频时长 - 根据
usage.input_image_count小于 5 免费 大于 5 收费 每张 1000 积分 - 音频暂时不收费
- 计费流程:先预扣 5 万积分、待生成后按实际消耗多退少补
- 扣费实例:
- 先提交任务 预扣
5w积分 - 获取任务状态.失败 返回
5w积分 - 获取任务状态.成功 得到 768p usage 实际消耗 8*2000+(7-5)*1000=20000 积分
json{ "task": { "usage": { "input_image_count": 7, "input_seconds": 3, "output_seconds": 5, "total_seconds": 8 } } }{ "task": { "usage": { "input_image_count": 7, "input_seconds": 3, "output_seconds": 5, "total_seconds": 8 } } }- 返还 :
5w积分-2w积分=3w积分
- 先提交任务 预扣
OpenAi-HK