15-second TV spot
A complete vertical commercial with paced product reveals.
Seedance 2.5 is ByteDance's AI video model for one-shot clips up to 30 seconds, synchronized sound, and output up to 4K. On Reviral, you create in your browser from text, images, video, or audio references, with the credit price shown before you render.
Render proof
These 6 real Reviral renders show product, action, testimonial, unboxing, and lifestyle formats from 15 to 30 seconds.
A complete vertical commercial with paced product reveals.
Fast action, camera movement, and a consistent subject.
A direct-response product moment built for the first 3 seconds.
A longer creator-led story with a clear spoken structure.
A try-on flow that keeps the product visible across the shot.
A lifestyle setup with controlled motion and visual continuity.
Direct answer
Seedance 2.5 is the latest multimodal video generation model from ByteDance, released July 31, 2026. It creates a continuous video up to 30 seconds from a written prompt, a starting image, reference video, reference audio, or a mix of those inputs. Its output includes synchronized dialogue, atmosphere, and sound effects in the same generation.
The model supports output up to 4K and as many as 50 combined references: 30 images, 10 videos, and 10 audio files. That reference budget keeps a product, character, location, movement, and soundtrack aligned inside one shot. Reviral puts the workflow in a browser: pick a look, add inputs, see the credit total, render, preview, and download from one workspace.
Three-step workflow
Choose a visual direction and an aspect ratio such as 9:16 for short-form feeds or 16:9 for widescreen video.
Write the shot in plain language, then add any of the 30 image, 10 video, or 10 audio reference slots you need.
Check the exact credit price, start the render, then preview and download the completed video in your browser.
One continuous shot
Seedance 2.5 generates up to 30 seconds in one shot, twice the 15-second ceiling of Seedance 2.0. Reviral also exposes shorter choices for tighter scenes. The proven 720p tier runs from 4 to 29 seconds, while the model's full duration set reaches 30 seconds. One-shot generation preserves motion and subject continuity without joining 2 separate clips.
30s
maximum one-shot length
Native synchronized audio
Yes. Seedance 2.5 generates dialogue, ambience, music cues, and sound effects inside the same render as the picture. Spoken words and visible action share one timeline instead of a separate audio pass. You can also supply up to 10 audio references to guide rhythm, voice, or atmosphere, alongside 30 images and 10 video references.
Transparent before render
The proven customer-visible 720p tier starts at 19 credits, equal to $0.37. Price scales with duration, and the exact charge appears before every render. The table lists only the visible tier; unproven tiers do not appear as public options.
Seedance 2.5 · 720p
Proven visible tier · 4–29 seconds
| Video length | Credits | USD equivalent |
|---|---|---|
| 4 seconds | 19 credits | $0.37 |
| 10 seconds | 48 credits | $0.94 |
| 15 seconds | 72 credits | $1.40 |
| 29 seconds | 139 credits | $2.71 |
USD equivalents use the current value of $0.0195 per credit and round to the nearest cent. Reference media does not add a separate multiplier to this tier.
Version comparison
Seedance 2.5 doubles the one-shot ceiling from 15 to 30 seconds and expands control through 50 combined references. Both versions generate sound; the 2.5 upgrade adds audio-reference input and stronger instruction following for dense prompts.
| Capability | Seedance 2.5 | Seedance 2.0 |
|---|---|---|
| One-shot duration | Up to 30 seconds | Up to 15 seconds |
| Output capability | Up to 4K | Up to 4K |
| Reference capacity | Up to 50 combined | Up to 15 combined |
| Audio references | Up to 10 | Up to 3 |
| Native sound | Yes | Yes |
| Prompt following | Improved complex-shot control | Earlier instruction model |
Four creator workflows
Build a 15- or 30-second product story with sound and a direct response structure.
UGCTurn 1 product image and a brief into creator-style product footage.
AVGuide a speaking subject with reference images, dialogue, and synchronized audio.
9:16Create 9:16 scenes for TikTok, Reels, and Shorts from a written script.
Eight direct answers
Seedance 2.5 is ByteDance's AI video generation model, released July 31, 2026. It turns text, images, video, or audio references into a video clip. Reviral runs it in your browser so you can render, preview, and download without installing software.
Seedance 2.5 generates up to 30 seconds of video in one shot, without stitching multiple clips together. Reviral also offers shorter lengths, and the visible 720p tier runs from 4 to 29 seconds. Longer clips use more credits.
Yes. Seedance 2.5 generates synchronized dialogue, ambient sound, and effects in the same render as the video. It also accepts up to 10 audio references, alongside image and video references, to guide timing and sound.
The visible 720p tier starts at 19 credits, equal to $0.37, for 4 seconds. A 10-second render costs 48 credits, equal to $0.94. 1080p costs more per second; 720p is shown here. Reviral shows the exact credit price before you start every render.
Seedance 2.5 extends one-shot video from 15 to 30 seconds, accepts up to 50 combined references, adds audio-reference input, and follows complex prompts more closely. Seedance 2.0 also generates sound; native audio is not new to 2.5.
Yes. Seedance 2.5 accepts up to 30 reference images, 10 reference videos, and 10 reference audio files. Add your product photos, character images, motion clips, or sound references beside the prompt before you render.
Yes. Seedance 2.5 is live on Reviral today. Sign in, open the creator, add a prompt or reference media, review the exact credit price, and start the render in your browser.
No. When a render fails to finish, Reviral refunds its credits automatically. A completed video uses the credits shown before the render starts, so the price remains clear before you create.
30 seconds · synced sound · up to 4K
Start with a prompt or your own reference media. See the exact credit price before 1 render begins, then preview and download in your browser.
Start freeLast updated: August 17, 2026