Newsletter
Join the Community
Subscribe to our newsletter for the latest news and updates
An upcoming multimodal AI video generation model with stronger control, consistency, and longer clips.
Seedance 2.1 is an upcoming multimodal AI video generation model by Kinovi, building upon Seedance 2.0. It aims to provide enhanced control, consistency, and longer video clips, moving towards production-ready creation.
createTask endpoint with a POST request. Include your API key in the Authorization header as a Bearer token. Specify the model as seedance2-1 and provide inputs such as a prompt, duration, resolution, aspectRatio, and optionally imageUrls, videoUrls, audioUrls, seed, upscaleResolution, callBackUrl, and mode.taskId returned from the createTask request to poll the recordInfo endpoint via GET requests. Continue polling every 2 seconds until the status is success or fail.output field in the recordInfo response will contain URLs to the generated video(s).Seedance 2.1 is expected to be available in late June to early July 2026.
The biggest upgrade is enhanced control, including richer multimodal input, longer videos, frame-level editing, better prompt adherence, smoother motion, and stronger character consistency.
Seedance 2.0 established cinematic 4K video generation. Seedance 2.1 builds on this with more context, precise editing, better instruction following, and improved consistency for longer or more complex videos.
Yes, Seedance 2.1 is expected to support videos up to 30 seconds.
Seedance 2.1 is expected to support up to 20 images, 6 audio files, and 6 video clips in one generation.
Yes, Seedance 2.1 is expected to offer more reliable character identity preservation across shots and longer scenes.
Creators, brands, agencies, and product teams needing more controlled AI video generation for product videos, social ads, character content, and reference-driven workflows.