ByteDanceVideo
Seedance 2.5
Seedance 2.5: ByteDance's most advanced video model: text, image and multi-reference to video with native audio.
- New
- Premium
At a glance
Seedance 2.5 by the numbers
- 15 s Max length per generation
- 1080P Max quality
- Audio Sound generated with the picture
- 7 Aspect ratios
- 2 Ways to create
- ByteDance Publisher
Use cases
What you can create with Seedance 2.5
Commercial spot
Hero shot, material close-ups and a final packshot in one video, with a synced soundtrack.
Action scenes
Sports, dance, chases, collisions: physics and complex motion stay believable.
Consistent character
Start from a character or product photo: it keeps its look from the first to the last frame.
Complex camera work
Tracking shots, dolly zooms, rack focus, handheld: describe them and Seedance executes them.
Gallery
Examples made with this model
Overview
What is Seedance 2.5?
Seedance 2.5, released in July 2026, is ByteDance’s most advanced video model. It generates picture and sound in the same space, so action and audio are natively in sync, and it handles real-world physics and director-level camera moves. ByteDance announces up to 30 seconds in a single shot, with no visible seams.
In Havincy, Seedance 2.5 produces 4 to 15 second clips from a text or an image, in 720p or 1080p. In image-to-video, it keeps your character’s or product’s look for the whole clip. It is the top-end model of the video catalog, best kept for shots that need to impress.
- Image, video and audio references
- Native audio
- Clips up to 15 s
- 720p and 1080p
Seedance 2.5: specifications
- Publisher
- ByteDance
- Generation modes
- text → video and image → video
- Available durations
- 4, 5, 6, 8, 10, 15 s
- Resolutions
- 720p, 1080p
- Aspect ratios
- auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
- Audio
- Sound generated with the video
- Available since
- July 20, 2026
- License
- Commercial use allowed
- Cost
- View pricing
The credit price depends on your settings and is shown before every generation.
Specs taken from the publisher’s public model page. See the documentation on fal.ai
Prompt examples
Prompts to try with Seedance 2.5
A drone shot gliding over a misty pine forest at sunrise
A chef flips a pancake in a sunlit kitchen, slow motion
A neon-lit street in the rain, a cyclist rides past the camera
Use cases
Compatible skills
Text → video
Clip vertical court généré depuis une description.
Image → video
Anime une image de référence en clip court.
Social post
Clip vertical prêt pour TikTok / Reels / Shorts.
Product demo
Clip qui met en avant l’usage ou le bénéfice d’un produit.
Cinematic scene
Plan de film narratif : mouvement caméra maîtrisé, lumière et mise en scène.
Trailer shot
Plan à fort impact type bande-annonce : tension, contraste, mouvement percutant.
How it works
How to create with this model
- Describe the desired result in a Havincy conversation.
- Add a reference when the style, product, or character needs to be preserved.
- Generate, then ask for a variation or improvement without starting over.
FAQ
Frequently asked questions about Seedance 2.5
Is Seedance 2.5 free?
No. There is no unlimited free access from the publisher; in Havincy, each generation uses credits from your plan. Test your idea first on Seedance 2.0 Fast or Kling 3.0, which cost less.
What’s new compared to Seedance 2.0?
Seedance 2.5 doubles the native duration (30 seconds at the publisher), accepts up to 50 references instead of 12, and generates sound in the same space as the picture for tighter sync.
When was Seedance 2.5 released?
Seedance 2.5 has been available on fal since July 20, 2026, and in Havincy shortly after.
What’s the maximum duration in Havincy?
You can pick 4, 5, 6, 8, 10 or 15 seconds, in 720p or 1080p, with native sound.
How much does Seedance 2.5 cost in Havincy?
The credit price depends on the settings you pick (length, quality, sound, number of images) and is shown before every generation. See the Pricing page for details by plan.
More models to discover
A starting point for your next creation.
Kling 3.0 Pro
Kling 3.0 Pro: Kling's flagship: cinematic video from text or an image, with native audio and clips up to 15 s.
Veo 3.1
Veo 3.1: Google's Veo 3.1: realistic video with synchronized audio, from text or an image, and extendable clips.
Seedance 2.0 Fast
Seedance 2.0 Fast: The fast Seedance 2.0: low-cost videos from text, images and references, with native audio.
Ready to create with Seedance 2.5?
Describe your idea in Havincy, then refine the result in the conversation.