Kling v3.0
Native audio-enabled video generation with Kling's 3.x motion language.
发布于March 2026
Wan 2.7 AI Video Generator
Wan 2.7 is Alibaba's latest video generation model with first/last frame control, multi-reference input, and instruction-based editing. Generate 5–15 second videos in 720P or 1080P.
内容政策提示
不符合政策的内容生成将失败——这不是诈骗或网站错误。Seedance 2.0 模型最近收紧了内容限制，所有真人面部输入（自拍、肖像、名人）现在在模型层面被阻止。暴力和 NSFW 内容也会被拒绝。您仍可以使用插画、动漫角色、AI 生成的面部或风格化艺术作品创作精彩的视频。
Core capabilities of Wan 2.7
What makes Wan 2.7 stand out
Wan 2.7 brings first/last frame control, multi-reference input, instruction-based editing, and extended 15-second output to Alibaba's video generation line.
First / Last Frame Control
Set the opening and closing frame. Wan 2.7 generates the motion between them — precise cinematic control without complex prompting.
Lock in both the starting and ending composition before generation begins.
Ideal for product reveals, character transitions, and scene cuts.
Removes the guesswork of achieving a specific end state.
Multi-Reference Input
Feed up to 5 reference videos simultaneously. The model uses them to guide character, environment, and motion style.
Upload up to 5 reference videos to shape the output.
Maintains visual consistency across character and scene.
Useful for brand content, fashion, and product continuity.
Instruction-Based Editing
Edit existing videos with natural language. Change backgrounds, lighting, clothing, or style without re-generating from scratch.
Describe changes in plain text — no timeline editing required.
Swap backgrounds, update clothing, or adjust lighting with a prompt.
Iterate fast without losing the core motion of the original clip.
Up to 15 Seconds
Generate clips up to 15 seconds — 3× longer than earlier Wan models. Enough for a complete product shot or scene.
Output at 5, 10, or 15 seconds depending on the scene.
Supports 480P, 720P, and 1080P resolution.
Works in 16:9 landscape and 9:16 portrait orientations.
Explore more models
Other Models to Try
Compare Wan 2.7 with other top video generation models available on this platform.
FAQs
常见问题
关于 Seedance 2.0 与平台
What is Wan 2.7?
Wan 2.7 is the latest video generation model from Alibaba Tongyi Lab, released in March 2026. It upgrades Wan 2.6 with first/last frame control, up to 5 simultaneous reference video inputs, 9-grid image input, instruction-based video editing, and improved motion physics.
What is first/last frame control in Wan 2.7?
First/last frame control (FLF2V) lets you set both the opening and closing frames of a video. Wan 2.7 generates the motion between them. This gives you precise cinematic control — set a start position and an end position, and the model interpolates the action.
How long can videos be with Wan 2.7?
Wan 2.7 supports 2 to 15 seconds per clip. This is a significant improvement over earlier Wan models, which capped at around 5 seconds. On this platform, you can generate 5, 10, or 15-second clips.
What modes does Wan 2.7 support?
Wan 2.7 supports text-to-video, image-to-video, first/last frame video (FLF2V), and instruction-based video editing. On this page, text-to-video and image-to-video are available.
What resolutions does Wan 2.7 support?
Wan 2.7 supports 480P, 720P, and 1080P output. Both 16:9 landscape and 9:16 portrait aspect ratios are supported.
Is Wan 2.7 open source?
Wan 2.1 was fully open source under Apache 2.0. Wan 2.7's open-source release status had not been officially confirmed at launch — check the Alibaba Wan GitHub (github.com/Wan-Video) for the latest.
How does Wan 2.7 compare to Wan 2.6?
Wan 2.7 adds first/last frame control, 9-grid multi-image input, up to 5 reference video inputs, and instruction-based editing — none of which were in Wan 2.6. Maximum duration increased to 15 seconds. Motion physics accuracy and character consistency were also improved.
还有问题？我们随时帮助