PixVerse/v6

PixVerse V6, released on March 30, 2026, is the most significant architectural upgrade of the PixVerse series to date. V6 has added more than 20 types of parametric movie-level camera movement controls, covering push rail, lift, surround, tracking, etc. Each type of camera movement can adjust the speed and easing curve. Native audio synchronization produces ambient sounds, sound effects and dialogue simultaneously with video generation, eliminating the need for post-production. The multi-camera engine supports defining 2 to 3 scene sequences in a single build and maintaining character and lighting consistency. V6 increases the video duration of 1080p native output from 8 seconds in V5.6 to 15 seconds, and supports five generation modes: Vincent, Picture, Transition, Expansion and Multi-shot. Multi-subject reference generation (r2v) supports inputting 2 to 7 images to intelligently fuse different subjects, ranking second in the global r2v track.

Context
0 tokens
Max output
0 tokens
Input
llm image video
Output
video

Pricing

Price: $0.005 / request

360p without : $0.05 / second

360p with : $0.07 / second

540p without : $0.07 / second

540p with : $0.09 / second

720p without : $0.09 / second

720p with : $0.12 / second

1080p without : $0.18 / second

1080p with : $0.23 / second

Endpoints

  • openai

More models

  • happyhorse-1.1-i2v
  • wan3.0-video

    Wan3.0-Video 是万相3.0系列的旗舰全能视频生成模型,采用 All-in-One 设计,无需切换模型即可覆盖文生视频、图生视频(首帧/首尾帧)和参考生视频等全部任务类型。其核心突破在于多模态素材理解能力:单次支持最多20个多模态素材参考,涵盖图片、视频、音频、PDF/PPT/DOCX文档及公开网页链接,模型可自动解析文档和网页内容并生成风格一致的新视频。生成时长大幅提升至最长30秒,帧率30fps,原生输出台词、BGM和音效,支持480P、720P、1080P分辨率输出。底层采用时空联合注意力扩散架构(Spatio-Temporal DiT),在隐空间内同步计算画面几何特征与帧间运动流,提升了长时序视频的稳定性。

  • dreamina-seedance-2-0-260128

    Dreamina Seedance 2.0 是字节跳动 Seed 团队开发的音视频生成模型,并集成到 CapCut 中。该模型采用双分支扩散 Transformer 架构,在生成链路中实现视觉与听觉信息流的并行处理,通过注意力/桥接机制实现帧级精度的声画同步。模型支持文本、图像、音频和视频多模态输入,最高可同步解析 12 个参考文件,生成长达 15 秒的视频,具备原生音视频同步生成能力。模型配备了隐形水印和 C2PA 内容凭证等保护措施。

  • dreamina-seedance-2-0-mini-260615

    dreamina-seedance-2-0-mini 于 2026 年 6 月 15 日发布,是 Seedance 2.0 系列中最快、最实惠的版本,主打高性价比路线,价格约为标准版的 50%。尽管定位更轻量,其生成效果与 Seedance 2.0 Fast 较为接近,在部分场景下已能满足日常生产需求,并专门针对电商和广告场景进行了深度优化。模型支持原生音频、摄像机控制和图像或视频输入,输出分辨率限于 480P 和 720P,不支持 1080P。该模型保留了核心的参考生成能力,用户可通过融合提示词与最多 12 个多模态参考素材锁定人物一致性。