Video model · by ByteDance
Turn a product photo or a prompt into a video ad with ByteDance's Seedance 2.5. Clips from 4 to 30 seconds, 480p to 1080p, with sound generated alongside the picture.
Seedance 2.5 is ByteDance's video generation model, announced in 2026 as the successor to Seedance 2.0. On Avocado AI you use Seedance 2.5 for text to video, image to video and reference to video, with clips from 4 to 30 seconds at 480p, 720p or 1080p and native audio.
Seedance 2.5 in the video workbench
Drop in a start frame, and an end frame if you want one. Describe the motion, pick the model, resolution and length, and get a product film.
Real Seedance 2.5 output. DAYFORM is a fictional brand.
For ecommerce brands
Animate your product photo into a 9:16 ad for Reels and TikTok, with sound.
Seedance 2.5 is one of the UGC studio engines: a creator look plus a script becomes a talking video.
Up to 30 seconds in one clip, enough for a full product story or demo.
Specs on Avocado
| Made by | ByteDance |
|---|---|
| Modes on Avocado | Text to video, image to video, reference to video |
| Durations | 4 to 30 seconds |
| Resolutions | 480p, 720p, 1080p, plus a 480p draft you can finalize at 1080p |
| Audio | Native audio |
| Example credits | 8 for 4s at 480p, 19 for 5s at 720p, 94 for 10s at 1080p |
| Plan access | Every paid plan |
Credits are the live workbench prices. The Generate button always shows the exact cost before you create. Plans in EUR, see pricing.
Brands use Seedance 2.5 for product video ads, UGC-style creator videos and short product films. Upload a product photo as the start frame, describe the motion and the setting, and get a clip with sound. Reference mode keeps a product, a person or a style consistent across shots.
Each clip runs from 4 to 30 seconds, chosen one second at a time in the video workbench. Longer clips cost more: 15 seconds at 720p is 57 credits and 30 seconds at 1080p is 281. For a longer commercial, join several clips in a Flow with Director mode.
Seedance 2.5 bills per second of output. A 4 second clip at 480p is 8 credits, 5 seconds at 720p is 19 and 10 seconds at 1080p is 94. Seedance 2.5 is on every paid plan, from EUR 19.99 a month for 100 credits. You see the price before you generate.
Draft mode renders a cheap 480p version first so you can check the motion and framing. If you like it, finalize the same draft at 1080p from the jobs panel or your library, without writing the prompt again. A draft can be finalized within seven days of creating it.
Updated September 26, 2026
Yes. Seedance 2.5 is live in the video workbench with text to video, image to video and reference to video, and it is one of the engines in the UGC studio. It is included on every paid plan and bills in credits by length and resolution.
Seedance 2.5 is ByteDance's newer model. On Avocado it runs up to 30 seconds per clip, while Seedance 2.0 runs up to 15 seconds. Seedance 2.0 is still available, including Fast and Mini variants and a 4K option for Growth and Pro plans.
Standard Seedance 2.5 may reject start frames that show real human faces. For those, Avocado offers Seedance 2.5 Faces, which accepts real faces as the start frame at up to 720p. It needs a Starter, Growth or Pro plan and a likeness consent confirmation.
Yes. Seedance 2.5 generates audio with the video, such as ambience, effects and voices that match the scene, so a clip can be posted without adding a soundtrack. You can still add music or a voiceover from Audio Studio if you want more control.
Yes. Every paid plan includes full commercial rights and no watermark, so you can run what you make in paid ads, on product pages, in email and on social. Results are saved to your Avocado library, ready to reuse in a video, a Flow or a Super Agent post.
Yes, through the Avocado MCP server. Connect it to Claude, Codex, Cursor or your own agent and ask for a video with a Seedance 2.5 model. Use dry run to see the credit quote first. The job runs on your Avocado credits and the video lands in your library.
More models
Google's fast image model for product shots and edits, up to 4K.
Google DeepMind's video model with native audio and start and end frame control.
Kuaishou's video model in Standard, Pro and native 4K, 3 to 15 seconds.
OpenAI's image model and Avocado's default, for product images and precise edits.
Seedance 2.5 and every other image and video model in one workspace, on one credit balance.
