PixVerseModel release

PixVerse · Model release

V6 Max
Control the moment. Keep the story coherent.

V6 Max is a directed video-generation model for cinematic storytelling, dialogue-driven scenes, music-led work, brand films, and creative shorts. This release covers text-to-video and image-to-video, with audio-video sync, outputs up to 15 seconds, and up to 1080p.

The source brief covers six capability areas and 21 release examples. Results vary by prompt, reference image, and scene.

Open PixVerse APIBrowse all release examples ↓

01 · Cinematic control

Direct the shot. Hold the world together.

The release brief opens with film-scale examples: action continuity, emotional performance, focus transitions, and controlled camera movement.

Cinematic action · 15 s · 16:9

Surreal city skateboard chase

高速移动+画面对象稳定

A moving camera, a stable subject, and a complete action path.

A high-speed skateboard journey moves from a rain-lit street through a series of jumps and transitions while the rider, environment, and camera direction remain coherent.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_视频以连续跟拍呈现一段超现实的城市.mp4

Full prompt from the release brief

视频以连续跟拍呈现一段超现实的城市滑板冒险:在雨夜霓虹街头踩着滑板疾驰。它压低身体穿过撑伞的人群,跃入斑马线,在出租车与其他车辆之间穿行,随后跳上狭窄隔离带,进入列车驶过的轨道区域,滑行时身后迸出火花。
接着,它腾空跃上高架道路,沿弯道快速滑行,又从道路边缘飞跃至旁边高楼,穿过敞开的落地窗进入室内,落地后转向镜头、逐渐停稳。镜头由低角度正面跟拍转为背后追随,湿润路面的灯光倒影、快速掠过的车辆与连续跳跃共同营造出强烈的速度感和游戏般的奇幻氛围。

Cinematic dialogue · 15 s · 16:9

Courthouse confrontation

情绪精准表达+精准镜头控制

Performance control carries the scene from accusation to resolve.

A woman confronts a man after a courthouse hearing; the model is directed through small emotional changes, eye-lines, breath, and a restrained push-in.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_Use_the_supplied_ (2).mp4

Full prompt from the release brief

Use the supplied image as the exact first frame. Create a 15-second, 16:9 cinematic realistic image-to-video scene in one continuous shot.

Maintain the faces, wardrobe, courthouse stairwell, cool daylight, and initial composition of the reference. The woman in the foreground is the emotional center. The man remains several feet away in the right background, softly out of focus. Her eyeline must connect naturally to his position; she never looks into the camera.

STORY AND PERFORMANCE
She once trusted this man to stand beside her when she needed justice. He failed to appear. Now she confronts him after a courthouse hearing.

She wants an explanation, but she is frightened that his answer will confirm he deliberately abandoned her. This conflict drives her performance: she searches his face for reassurance even while accusing him.

Her emotion develops through disbelief, exposed hurt, a brief loss of breath, and finally quiet resolve. Convey these changes through small, motivated actions rather than exaggerated crying. Do not stack every facial gesture at once.

0–2 seconds:
Begin with her mouth closed, matching the reference.
She watches him, waiting for him to speak first.
Her lower eyelids tighten slightly. She takes a shallow breath and swallows before finding her voice. He remains silent.

2–5 seconds:
In a low, strained voice, she says:
“I kept looking at the door. You never came.”

The first sentence carries a trace of hope remembered.
Before “You never came,” her gaze searches his face.
The accusation emerges as a painful realization, with a slight break on “came.” No shouting.

5–8 seconds:
She tries to continue but cannot immediately draw enough breath.
Her lips part; a small, involuntary breath escapes.
She looks down at the folder for a moment and tightens her fingers around its edge. One tear slips down her cheek.
She is embarrassed by losing control and tries to steady herself.
No dialogue during this beat.

8–12 seconds:
She raises her eyes toward him again.
Her jaw relaxes slightly; her voice becomes quieter and more deliberate.
She says:
“We promised we’d stand up for people.”

On “we,” a faint lift of her eyebrows asks whether he remembers being that person. Her eyes remain wet. Her expression is wounded, not threatening.

12–15 seconds:
The camera completes a very gentle push toward her face.
She holds his gaze and asks softly:
“Why was I different?”

A slight hesitation before “different” reveals how personal the wound is.
Finish the line by approximately 14 seconds.
Hold the remaining moment on her waiting for an answer, lips barely trembling. Do not resolve the confrontation.

THE MAN
He never speaks or mouths words.
At her final question, he almost raises his eyes, then fails to meet hers.
His guilt is expressed through stillness and avoidance, without dramatic gestures. Keep him out of focus.

CAMERA
One continuous eye-level shot.
Keep the woman’s eyes sharply focused throughout.
Use only subtle handheld breathing and a slow, nearly imperceptible push-in. No cuts, reverse angles, focus pulls, or sudden zooms.
Preserve spatial continuity as she turns her gaze toward him.

AUDIO
Natural American English, synchronized to her lips.
Her speech must sound physically connected to her breathing: a tight throat, uneven exhalations, and one brief voice break.
Keep the words intelligible. Do not add sobs to every phrase.
Only quiet stairwell ambience, breathing, and slight clothing movement.
No music, voiceover, male dialogue, or additional speech.

VISUAL STYLE
Restrained independent drama, cool soft window light, muted stone and olive tones, natural skin texture, shallow depth of field, subtle film grain.
Preserve facial identity and realistic anatomy.
No beauty filter, theatrical grimacing, continuous screaming, subtitles, on-screen text, or watermark.

Atmosphere · 13 s · 16:9

Rainy-night detective

主题氛围表达,跨镜头人物一致性稳定

A single take where atmosphere, focus, and continuity work together.

A detective pauses beside a rain-streaked window as the camera moves laterally, shifts attention to the reflection, then returns to the subject.

Prompt & video breakdown

Source video: PixVerse_mp4_media_web_ori_9391effe-98fa-43b3-89c9-587d85fff139_seed0.mp4

Full prompt from the release brief

13秒雨夜悬疑片段,沿首帧窗边侦探的中近景起镜,65mm长焦观感。相机在室内,与窗面呈斜角,侦探脸、玻璃反射与窗外人形成三层空间。
0–4秒,杯沿抬至嘴边却停住,他先向右移眼球,下颌再轻轻收紧,头不动。焦点保持近侧眼睛,玻璃上的雨滴略虚,霓虹不污染整张肤色。
4–8秒,摄影机向右平移约半米,维持侦探在左三分线,利用视差让风衣人倒影进入右侧负空间;用约一秒半的拉焦,把注意力从眼睛移到窗上倒影。实体与反射服从同一窗面几何,雨水在玻璃前层持续下滑。
8–11秒,杯子落桌轻响,镜头停止侧移,焦点返回侦探。他这才转头向窗外,风衣人沿既有方向走出窗边,不能凭空溶解。
11–13秒停在人物半侧脸近景,空出的窗面仍有霓虹,怀疑留在视线里。单镜头完成,曝光稳定,表演靠杯子的停顿与眼动,不放大成惊恐。

Performance detail · 14 s · 16:9

Driver's-seat focus shift

自然转焦能力

A small gesture becomes a readable focus transition.

A tired bus driver shifts focus from her eye to a returned toy and back again, with small facial and hand actions carrying the change.

Prompt & video breakdown

Source video: PixVerse_mp4_media_web_ori_ebf675ee-1734-4d1f-8af9-3e6513891473_seed0.mp4

Full prompt from the release brief

凌晨公交车驾驶席,一名50岁女司机,蓝色工服,窗外路灯逐个掠过;车辆停靠时的柔和仪表光。
65mm、T2.8侧面近景、24fps,驾驶席机位固定;仪表板低位冷光与路灯暖边光不过分夸张。乘客手进入时轻移焦到玩具,摆正后回到司机眼睛;揉眼时皮肤发生局部挤压,松手后恢复,疲惫的表情没有磨皮。
0–3秒:固定侧面近景,车已停稳。
3–7秒:司机揉一次眼角,整理衣领。看向车门处,一位乘客的手把忘带的小玩具放到仪表台。
7–12秒:她先意外抬眉,后露出疲惫的暖笑,伸手把玩具摆正。
12–14秒:最后看回前方。
乘客的手放下玩具后退出画面,她把玩具摆正,笑意先在疲惫的眼周浮现,再牵动嘴角。怠速、衣领摩擦。

Film look · 14 s · 16:9

Underwater cathedral

电影美学

Depth, light, fabric, and movement hold across a long take.

A figure rises through an abandoned underwater church as light, fabric, hair, and suspended particles move with the resistance of water.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_从首帧接续一段14秒写实水下长镜头.mp4

Full prompt from the release brief

从首帧接续一段14秒写实水下长镜头,维持青绿深水、乳白薄裙、金色斜射光束与古老石材的克制色彩。摄影机在女主左后方胸部高度,以35mm视角缓缓向前漂移;她双脚轻触教堂石阶,先抬头看向右上方破损天窗,嘴边逸出两三颗气泡,头发在水中缓慢展开。最初4秒不急着展示大场面,让光束里的细小悬浮颗粒从她脸旁经过。随后她轻蹬石阶,双脚离地,身体缓慢向上浮升,手臂自然张开以稳定姿态,裙摆晚半拍才离开石阶。镜头向右绕过她的侧面,同时稍稍后退,逐渐显露高耸柱廊、残破长椅和上方洒下日光的巨大圆窗;人物仍处在原来的中央过道,不能突然换到另一座建筑。最后5秒,镜头与她同速上升,从侧面中景靠近至肩部近景,一道窄光从额头滑到眼睛,她原本绷紧的眉心慢慢松开,目光安静地追随上方亮处;裙摆和长发继续在身后滞后漂动,气泡从前景向上离开。运动要有水的阻力,镜头与人体都缓起缓停;水下能见度足够辨认结构,保留深处暗部,不把水拍成空气或浓烟,不出现字幕与对白。

02 · Brand storytelling

Make the idea legible from first frame to final beat.

The source brief includes beauty, TVC, and creator formats as capability proof for visual identity, generated copy, and performance control. This is a model release surface, not a product catalog.

Brand film · 8 s · 9:16

Cocoa Soft Glow

商品跨镜头文字一致性稳定

Packaging detail and generated copy stay legible through the shot.

A beauty sequence moves from unboxing to hero close-up and final packshot while label, color, ingredients, and generated copy stay consistent.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_创作一段高质量、逼真的_8秒竖屏_.mp4

Full prompt from the release brief

创作一段高质量、逼真的 8秒竖屏 9:16 比例护肤品商业广告。准确保留虚构品牌“Cheng Guang Cocoa Soft Glow”的瓶身设计、标签、颜色、包装及整体构图。
场景 2 (0–2s):一位女性的双手轻轻从盒中拿起 Cheng Guang 瓶子。缓慢自然的手部动作,逼真的皮肤纹理,柔和温暖的灯光,营造高端的开箱体验。
场景 3 (2–4s):英雄式产品特写。瓶子直立放置,周围环绕着可可豆、可可脂和新鲜绿叶。镜头从左向右缓慢移动,呈现优雅的景深效果。
场景 4 (4-5s):极致微距特写,展示顺滑乳液质地被轻柔涂抹于肌肤之上。以逼真的微距细节展现奶油般轻盈的质地和光泽保湿效果。
场景 5 (5–7s):一只手优雅地握住瓶子并缓慢转向镜头。通过高端美容广告灯光,突出“Cocoa Butter and Serum-in-Lotion”(可可脂与精华乳液)的概念。
场景 6 (7–8s):最终英雄式镜头,展示被可可豆和绿叶环绕的 Cheng Guang Cocoa Soft Glow 瓶子。镜头缓慢推进,伴随微妙的闪光特效,奢华温暖的背景,聚焦产品的干净构图。
风格:逼真摄影,高端美容广告,电影级布光,真实手部动作,流畅镜头转场,浅景深,柔焦散景,暖金色调,4K 分辨率,高细节,自然动态,无畸变,无多余产品,不改变瓶身标签或品牌标识。

TVC · 15 s · 16:9

Emerald fragrance TVC

TVC广告质感,文字生成能力

A finished brand-film language with a controlled final frame.

A luxury fragrance sequence moves from talent and spray action to a stable final packshot with the line “ELEGANCE, DISTILLED.”

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_15_秒,169_画幅的奢华香水广.mp4

Full prompt from the release brief

15 秒,16:9 画幅的奢华香水广告。写实电影感时尚大片,极简奢华风格,浓郁的祖母绿环境,高对比度布光搭配优雅的柔和金色点缀,光泽反射与高端微距产品摄影
0–3s — 主角开场:
立即以清晰可见的女性形象开始,她位于画面中央,摆出自信的时尚姿势,身穿祖母绿丝绸连衣裙,手持 香水瓶置于胸前。从第一帧开始,必须清晰展示她的脸部、连衣裙和瓶子。背景为浓郁的祖母绿色,带有柔和的高光。明亮、高端且优雅。无暗部阴影、灰色背景、空镜或淡入效果。
3–6s — 喷洒动作:
平滑过渡到电影感特写镜头。她举起香水瓶,向颈部喷洒一次。细腻的香水雾气捕捉到金色的光线。她放下瓶子,轻轻触碰锁骨。动作优雅克制,具有真实的反射效果和雾气细节。
6–9s — 时尚动态:
切换至更宽的景别。她缓慢转身并向前走,同时摄像机向后追踪拍摄。她的祖母绿丝绸连衣裙和头发自然摆动。祖母绿氛围在她周围扩展,伴有细微的雾气和流动的金色光带。
9–12s — 产品高光:
过渡到香水瓶在祖母绿环境中漂浮的高端微距镜头。瓶子缓慢倾斜并捕捉金色高光。展示真实的玻璃质感、液体反射、细微的蒸汽轨迹和精致的琥珀色颗粒。保持瓶身几何形状完美稳定。
12–13.5s — 最终模特镜头:
切回紧凑的时尚镜头。她带着自信优雅的表情轻轻转头面向镜头。头发和丝绸面料在空气中轻柔飘动。金色轮廓光勾勒出她的剪影。
13.5–15s — 最终包装展示:
切换至干净的高端包装展示镜头。PERFUME香水瓶居中置于深黑色背景前,伴有祖母绿光晕和精致的金色轮廓光。显示文字:
“ELEGANCE, DISTILLED.”
一位精致的女声清晰地念道:“Elegance, distilled.”

Creator format · 14 s · 9:16

Creator-style beauty demo

UGC真实感

A natural creator performance with a clear spoken point.

A creator-style beauty clip keeps the person, visual identity, home setting, and spoken recommendation consistent across a fixed phone-camera shot.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_生成14秒916真人UGC电商短视 (2).mp4

Full prompt from the release brief

生成14秒9:16真人UGC电商短视频。严格以输入图为首帧,保留同一成年人物、商品外观、服装及家居环境。固定手机支架视角,日常自然光,轻松像向朋友推荐,动作适度,不用棚拍运镜。
动作与节奏:0–4秒,说第一句。4–9秒,用右手无名指围绕已有色点轻拍三次,色点随接触局部晕开。9–14秒,手离开脸颊,轻转头展示,说第二句。
第一句英文原文:"I start with one tiny dot of this peach blush."
第二句英文原文:"Tap it out, and keep the rest of the look simple."
只说以上英文,不翻译、不加其他台词;自然口语美式英语,音量清晰,嘴形同步,动作段保留自然停顿。只由画面中的人说话,结尾留约一秒停稳。连续单镜头,保持商品颜色、材质、数量和结构;物体只在实际接触后移动。保留轻微环境声和对应操作声,无配乐、无字幕、无价格或促销字卡。

Creator format · 14 s · 9:16

Creator-style beauty demo · alternate

人宠+商品互动

A second creator setup for testing continuity and delivery.

A second creator-style clip tests a different person and setting while preserving a direct-to-camera recommendation and visual continuity.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_生成14秒916真人UGC电商短视 (5).mp4

Full prompt from the release brief

生成14秒9:16真人UGC电商短视频。严格以输入图为首帧,保留同一成年人物、商品外观、服装及家居环境。固定手机支架视角,日常自然光,轻松像向朋友推荐,动作适度,不用棚拍运镜。
动作与节奏:0–4秒,说第一句。4–9秒,将开口缓慢移到猫鼻前,轻挤使肉泥从开口少量冒出,猫舔食。9–14秒,保持距离与握法,看镜头说第二句。
第一句英文原文:"A tiny treat for my very opinionated little roommate."
第二句英文原文:"I think that is a yes."
只说以上英文,不翻译、不加其他台词;自然口语美式英语,音量清晰,嘴形同步,动作段保留自然停顿。只由画面中的人说话,宠物不说话。结尾留约一秒停稳。连续单镜头,保持商品颜色、材质、数量和结构;物体只在实际接触后移动。保留轻微环境声和对应操作声,无配乐、无字幕、无价格或促销字卡。

03 · Dialogue and continuity

Give every character a role, a reaction, and a reason to stay in the scene.

Multi-person staging, dialogue assignment, turn-taking, lip synchronization, and prop continuity are all represented in the source release brief.

Short drama · 15 s · 9:16

The half-jade reveal

Dialogue and continuity

Let the reveal land through performance, props, and turn-taking.

A two-person historical drama builds from accusation to recognition while preserving the pouch, pendant, character positions, and spoken turns.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_沿用首帧的古代正厅、左侧绛红衣贵妇.mp4

Full prompt from the release brief

沿用首帧的古代正厅、左侧绛红衣贵妇、右侧青灰衣年轻男子,以及男子手中的系绳红布小袋,人物均为成年人。50mm平视双人中景,冷色门光勾边,烛光暖亮面孔。0–4秒,男子猛然向前一步,肩背绷紧,怒声问:“当年,你为什么丢下我!”贵妇原本严厉的目光微动,仍强撑着不退。4–8秒,男子停住,用空着的手拉开袋口,持袋的手把袋中唯一的半块白玉坠倒入另一只手的掌心,抬至胸前;镜头贴近到手与两张脸同框的中近景,焦点先落玉坠,再移到贵妇眼睛。8–12秒,贵妇看清缺口后瞳孔放大,嘴唇剧烈发抖,呼吸骤断,含泪喊:“我的孩子!”她伸出的手停顿一下,才轻轻托住他捧玉的手腕。12–15秒,男子的怒意瞬间瓦解,眉心上提,眼眶湿润,难以置信地说:“你……是我娘?”镜头缓推到侧向双人近景,贵妇一边哭一边急切点头,男子肩膀缓缓塌下。情绪幅度要大,哭泣来自呼吸、喉部和唇颤,不靠扭曲脸形;红袋始终留在首帧持袋的手,半玉始终留在另一只手,归属稳定。只出现半块玉,不补全、不发光;切镜保持贵妇左、男子右,保留烛火环境声和原声对白,无音乐盖住停顿,无字幕。

Short drama · 13 s · 16:9

Office identity reversal

Dialogue and continuity

A multi-character reversal with clear roles and reactions.

A workplace confrontation turns into an identity reversal through role assignment, focus changes, and distinct reactions from three characters.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_以输入图为第一帧,拍一场13秒的职 (2).mp4

Full prompt from the release brief

以输入图为第一帧,拍一场13秒的职场身份翻盘。左侧经理坐着,右侧穿米白西装的年轻女性站着,后方门口的银发男高管已经在场。用40mm平视双人中景开场,冷白窗光与木墙暖色反光相抵。经理突然拍桌,镜头快速正面聚焦到经理,经理身体前探,冲女子怒吼:“你被开除了!”镜头随拍桌短促推近,不做无来源的抖屏
女子先抿住嘴角、鼻翼一紧,吸气后抬眼,压住怒意。后方高管快走两步停在女子身后,向女子微微欠身,对着女子说:“总裁,会议准备好了。”焦点从经理转移到高管,再切回经理的同侧近景:他的怒容先定住,眉毛抬起,眼神在两人之间急移;他慌忙起身,椅子被腿向后推开,声音骤软:“总裁?”最后三秒切回中景,女子抬起下巴直视他,冷笑一声,说:“现在,谁走?”镜头低幅度缓推,保留经理僵站的局促姿态。强化怒斥、错愕、讨好与压场的反差,不能让三人同时张嘴;经理始终在左、女子在右,高管不穿过两人。人物面孔、桌上单只水杯及椅子位移跨镜连续

Everyday drama · 15 s · 9:16

Dinner and good news

多人同框台词和人物对应关系准确

Keep the conversation anchored to people, props, and positions.

A father and daughter share a simple dinner exchange while the envelope, bowls, seating positions, and dialogue remain connected across the cuts.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_父亲看门边女儿,说:“菜快凉了。”.mp4

Full prompt from the release brief

父亲看门边女儿,说:“菜快凉了。”身体稍向空椅倾斜。女儿低头松开鞋带,随后走到桌边,镜头沿桌边侧移,保持父亲在左、女儿在右。女儿说:“我今天拿到工作了。”把信封放到两人之间。信封触桌后切父亲近景,女儿肩留右前景;父亲停止咀嚼,愣半拍,眼角放松,逐渐露出收着的笑。随后切回同侧双人中景,父亲拉开右椅,说:“坐,先吃饭。”女儿坐下,末秒两人短暂对视。信封、饭碗与菜跨镜头位置一致,口型对应对白。

Dialogue continuity · 15 s · 9:16

Apothecary mix-up

跨镜头保持稳定人物一致性

Continuity can carry a small comic beat across the cut.

A comic apothecary exchange keeps two colored medicine bags, hand contact, character positions, and dialogue aligned through the action.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_书生伸手拿起左侧红绳药包,说:“这.mp4

Full prompt from the release brief

书生伸手拿起左侧红绳药包,说:“这一包给我。”镜头稍前推。医师先看包再看书生,左手按住他手腕,说:“那是给猫的。”接触后书生停手,两人静止半秒。书生低头看包,笑容逐渐僵住,再沿来路把红绳包放回左侧,完全松手后转向右侧拿起蓝绳包。镜头轻俯,持续看清双手与两包。医师偏头忍笑,书生回看后笑出来。最后两秒镜头停稳。连续拍摄,红蓝绳色、药包数量和位置一致,台词与各自口型同步。

04 · Music-led generation

Let image, movement, and sound share one timeline.

The release brief covers beat-aware visual changes, original music, singing, lip synchronization, and multi-performer timing.

Music rhythm · 12 s · 16:9

Synthesizer rhythm study

高度适配音乐节奏,音画同步

Let visual changes follow the structure of the music.

A musician performs to an original 120 BPM, 4/4 electronic track while light rings and camera movement change on the beat.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_原创120_BPM、44拍电子器乐.mp4

Full prompt from the release brief

原创120 BPM、4/4拍电子器乐,从0秒第一拍开始。0–4秒右手捏住旋钮顺时针转四分之一圈,滤波音色随之逐渐变亮;身后单一光环每半秒随拍短暂增亮再恢复,不复制。4–8秒双手按住合成器键区演奏,镜头向右弧移约45度,圆环、人物和倒影呈连续视差。8秒强拍,光环沿圆周均匀分成十二块,保持圆形分布。8–10秒十二块各自原地转半圈,10–12秒沿原位置重新闭合成完整圆环。人物和桌子保持实体,无人声、文字或额外乐器。

Vocal performance · 12 s · 16:9

Singer in the rain

嘴形高度同步音轨

Lip movement, breath, and melody stay connected.

A singer performs an original melody in the rain with mouth shapes, breath, microphone, and accompaniment directed together.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_0–2秒,歌手抬麦至嘴前并吸气,镜.mp4

Full prompt from the release brief

0–2秒,歌手抬麦至嘴前并吸气,镜头缓慢推近。2–6秒,用清晰普通话原创新旋律唱“雨停以后,我还在这里”,嘴形跟随每个音节,肩部自然呼吸。6–8秒闭口换气,麦克风仍在嘴前。8–11秒唱“等一束光,照亮你”,11–12秒闭口收尾,镜头停在胸像。全片120 BPM、4/4拍原创流行伴奏,第一拍位于0秒,换气处伴奏连续。轻雨落下且沿脸与衣物自然滑动。单镜头,人物面貌、耳饰与麦克风形状保持一致,无字幕。

Band performance · 13 s · 16:9

Band performance MV

多人场景下仍然保持高度契合音乐

A band performance built around one continuous musical structure.

A three-person band performance follows one original instrumental structure, with cuts that preserve instrument positions and shared timing.

Prompt & video breakdown

Source video: PixVerse_mp4_media_web_ori_87c0877e-b2a0-4eec-8501-cde29f0410dd_seed0.mp4

Full prompt from the release brief

乐队新单曲的演奏版MV要建立真实而有能量的乐手形象,适合音乐平台首发与现场风格宣传。采用每分钟120拍的原创摇滚器乐,以吉他乐句、鼓过门和最后齐奏串起三人配合,让观众感到他们正在共同演奏。
旧仓库舞台上,左侧黑衣吉他手、中后方红鼓组鼓手、右侧灰衣贝斯手站位明确。
35mm全景、T4,暖侧光穿过薄雾,冷轮廓光分开人物与背景,24fps。
0–4秒:35mm全景低速侧移,三人开始同一乐句;吉他拨弦、贝斯右手拨奏和鼓槌击打落在共同节拍上,三人互相看一眼。
4–8秒:第9拍切85mm吉他手双手特写,T4,左手在琴颈换把、右手拨弦;焦点沿指板短距离移动,琴弦位置、品丝和手指接触清楚。
8–12秒:第17拍切50mm鼓手中景,完整显示双手与军鼓、镲片;最后一小节完成短过门,鼓槌先接触鼓皮再回弹,镲片击中后持续振动。
12–13秒:切回35mm全景,三人齐奏一个收束重音;吉他和镲片余音延续,演奏者身体缓慢停下。
乐器、背带与鼓件保持固定结构;切镜不改变三人位置或演奏乐句。声音由同一首原创器乐连续承接,击打与可见接触同步,不添加观众、歌词和可读标志。

05 · Creative worlds

Move beyond realism without losing the action.

Stylized animation, cel shading, fantasy effects, martial-arts action, and paper-cut motion show how V6 Max can preserve a chosen visual language.

Stylized short · 10 s · 16:9

Ink-wash duel

水墨画风格,格斗动作稳定.

Make the visual language part of the motion.

Two sword fighters move through a misty bamboo grove as bodies, blades, smoke, and impact marks resolve as ink lines and washes.

Prompt & video breakdown

Source video: 短漫剧1.mp4

Full prompt from the release brief

薄雾竹林以流动水墨线条和淡彩晕染呈现,长袍剑客滑步向前,身后拖出逐渐溶进雾中的笔触;蒙面刺客后退时身体碎成墨滴再重新聚拢。剑客横斩,刀光化作粗黑墨线,墨点悬浮后晕散;刺客侧身闪避,留下灰色拖影,并用匕首刺出锯齿状墨痕。剑客低身旋转,宽袖像书法长锋般展开,金属交击令层层墨晕在空中绽放。终结一剑划过后,刺客化为铺开的墨池渗入地面雾气;剑客收刀,刃边墨烟细细卷起,竹林恢复清寂。

Animated action · 15 s · 16:9

Cel-shaded mech launch

赛璐璐风格

Keep a two-dimensional design system stable through action.

A pilot launches a cel-shaded mech while cockpit controls, hatch movement, exterior steps, and direction stay consistent.

Prompt & video breakdown

Source video: PixVerse_mp4_media_web_ori_924c4f48-1909-4526-a636-7bd91044b661_seed0.mp4

Full prompt from the release brief

保持二维机甲线稿与赛璐璐色阶。0–4秒,驾驶员拉回右操纵杆,半开舱门沿原导轨闭合,舱外亮光逐渐收窄,仪表蓝光映到脸上。4–7秒,镜头稍推近,她确认前方后推左杆;杆到位后机体轻震,头发与肩部稍滞后。7–12秒,切机甲外部低位全景,右脚向画面右方跨前,关节依结构转动,脚落地后液压收紧、重心移交,尘土从接地点扩散,镜头短震一次。12–15秒,切回原舱内近景,她握稳两杆并说:“出发。”内外红白机体身份、闭舱状态与行进方向一致。

Fantasy action · 15 s · 16:9

Arcane arena fight

高速打斗特效

Effects, bodies, and camera energy share the same visual world.

A magic duel combines spell effects, debris, camera energy, and character action inside a coherent fantasy arena.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_紫色暮空下的残破石质竞技场里,长发.mp4

Full prompt from the release brief

紫色暮空下的残破石质竞技场里,长发女术士抬起双掌,脚下多层法阵逐圈点亮,银色符文沿手臂攀升。持黑焰法杖的角族对手砸地升起火墙,女术士投出白色雷矛,电光撞上火墙后分裂成蛇形支流,在石面留下焦痕。她拖着霜雾低身冲近,冻结半空碎石,再以符文膝击击碎屏障。对手召出螺旋黑曜石刺,她在尖刺间翻跃,贴身双掌释放白光爆发,将角族击飞;黑焰、石块与晶尘悬在暮色中,余光缓缓消退。

Wuxia action · 15 s · 16:9

Wuxia sword exchange

武侠风打斗

Fast choreography stays readable as the fighters exchange positions.

Two sword fighters exchange attacks, change position, and resolve into a quiet back-to-back pose without a declared winner.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_开场,两人已经处于交锋状态。斗笠剑 (1).mp4

Full prompt from the release brief

开场,两人已经处于交锋状态。斗笠剑客跨步挥剑,剑锋横向扫过,带出橙红色的光影轨迹;浅衣剑客压低身体,侧身闪避,同时伸剑反击。双方随即起身逼近,接连使出直刺、横斩与抬剑格挡,脚步不断进退交错。宽大的袖口、袍摆和发带随着转身甩开,脚下也扬起阵阵尘土。
中段,两名剑客连续旋身换位,攻防节奏加快。镜头先以侧面全景展示完整的身法,再随着交锋拉近,突出近身擦过的剑刃与翻动的衣袖。一阵浓厚烟尘短暂遮住两人,只留下模糊身影;他们随即重新显现,在贴近身体的距离内挥剑拆招,剑刃相碰时迸出明亮火星。
随后,浅衣剑客跃起,双脚短暂离地,衣袍在空中展开;斗笠剑客则保持低重心,举剑迎击。落地后,两人再次拉开身位,又迅速前冲,剑锋在胸前和头顶交错。几次幅度较大的挥剑带起弧形暖色光痕,与逆光中的烟尘融在一起,让动作带有略微夸张的武侠特效感。
最后一轮交锋中,两人迎面逼近,挥剑后擦身而过,最终背对背停住。双方都将剑尖斜垂向地面,衣摆和发带仍在轻轻飘动,画面从密集攻防转为静止对峙,没有明确交代胜负。

Paper-cut animation · 15 s · 16:9

Paper-cut shadow fable

剪纸风小短片

A delayed shadow becomes a simple story about synchronization.

A paper-cut deer and its delayed shadow gradually synchronize after a small conflict, using motion and rhythm rather than dialogue.

Prompt & video breakdown

Source video: PixVerse_V6_Image_Text_1080P_剪纸寓言短片借慢半拍的影子讲一次从.mp4

Full prompt from the release brief

剪纸寓言短片借慢半拍的影子讲一次从不适应到彼此协调的相处。先让延迟形成小小冲突,再用鼻尖接触后的同步完成和解,适合以动作和音乐传达情感的动画栏目。
多层纸片剪纸动画,红色纸片鹿与深蓝色影子鹿,黄色月亮、米白树干,侧光突出纸层厚度。
红纸鹿在左,深蓝影子鹿落后半个身位,分层树干与月亮分居前后;正交侧视中远景,侧光显纸边厚度,24fps时间线、12fps关节姿态。
0–3秒:侧视横移跟随红鹿,前树视差大于月亮,蓝鹿姿态延迟12帧。红鹿向右走,蓝影子落后半身,动作晚半秒。
3–6秒:镜头随红鹿停下,蓝鹿再补一步碰上,纸层遮挡顺序稳定。红鹿停下,蓝影子又向前一步轻碰红鹿,红鹿回头。
6–10秒:固定全身中景记录红鹿抬蹄,蓝鹿严格晚半秒再做相同动作。红鹿抬起前蹄再落下,影子每个动作都晚半秒。
10–15秒:镜头轻推不改变正交关系,鼻尖接触后延迟逐渐归零,末段同步抬头。红鹿转身轻触影子鼻尖,二者从错拍逐渐同步,最后一起抬头看月亮。
纸鹿的鼻尖相碰,先前落后半秒的动作开始同步;剪纸边缘和关节清晰,横移显露前后纸层的轻微视差。纸张摩擦和轻节拍。

V6 Max core capabilities

Six areas of model improvement, stated plainly.

The source brief describes the release as a focused update for film, short drama, advertising/brand storytelling, music video, and creative-short production.

Instruction following and precise control

Respond to complex prompts and multiple constraints across subject behavior, emotional change, camera movement, shot design, composition, and transitions.

Realism and cinematic aesthetics

Improve lighting, materials, spatial depth, facial detail, wardrobe texture, color, composition, depth of field, and the overall finish of a scene.

Cross-shot consistency

Preserve character identity and object detail as scale, angle, viewpoint, and scene conditions change.

Music-rhythm synchronization

Align actions, visual changes, effects, and cuts with the beat, musical structure, and changing intensity of a track.

Multi-person dialogue and lip sync

Assign lines to the correct character, control turn-taking and pauses, coordinate mouth movement with speech, and preserve listener reactions.

Object consistency and text rendering

Improve the stability of product form, color, packaging, key design features, and generated brand or packaging text across shots.

Scope of this release

V6 Max is the latest V6-series version described in the source brief. The update covers T2V and IT2V only; other capabilities are outside this release scope. Use the prompts below as the complete source-backed direction for each example.

Source-backed release facts

A focused release surface with clear boundaries.

The overview, capability summary, prompts, and examples on this page are drawn from the Lark release document, accessed October 16, 2026.

2 modesText-to-video and image-to-video
15 sMaximum source-stated generation length
1080pMaximum source-stated resolution

Source: “V6 Max 模型介绍及场景效果”, Lark document revision 1012. The page includes all 21 source video examples and their prompt directions; current access and commercial terms should be confirmed on the live platform.

Release access, pricing, and scope

Start with the model surface that fits the shot.

V6 Max is available through PixVerse API access. The pricing table and promotion terms below reproduce the source brief; verify current availability, account requirements, documentation, and commercial terms before production.

  1. Choose text-to-video or image-to-video; both modes use the same source-stated rate table.
  2. Write the shot: performance, camera, continuity, sound, visual language, and any reference-image constraints.
  3. Confirm current access, limits, promotion status, and commercial terms before production.
Open PixVerse API

V6 Max stated rate table

For T2V and IT2V, the source brief states the same per-second rates for silent and audio-video generation.

ResolutionSilent · CNY/sAudio · CNY/sSilent · credits/sAudio · credits/s
360p0.1760.26469
540p0.2640.323911
720p0.3230.4401115
1080p0.6450.8212228

Platform and documentation links

The source brief lists domestic and overseas API platforms and documentation surfaces.

Domestic API platform ↗

Overseas API platform ↗

Domestic API documentation ↗

Overseas API documentation ↗

Two-week promotion in the source brief: API platform users receive the V6 per-second credit rate for two weeks from V6 Max launch; enterprise customers receive two weeks from their enablement date.

The source brief points to a separate PixVerse video-generation price lookup table for detailed pricing. This page reproduces the V6 Max table stated in the release document and does not replace the live commercial source.

In scope

Text-to-video and image-to-video only.

Output ceiling

Up to 15 seconds and 1080p, with audio-video sync described in the brief.

Prompt surface

The full source prompt for every release example is available inside its video card.

Commercial check

Verify live account access, limits, promotion status, and terms before production.