In shortSeedance 2.5 Flash is a faster, cheaper tier of Seedance 2.5 on BytePlus ModelArk (model id dreamina-seedance-2-5-flash-261115). It lists at US$7.0 per million tokens without video input, about 35% below Seedance 2.5, and outputs 480p or 720p clips of 4 to 30 seconds. In our one test it finished in 76 seconds against 139, with flatter detail.
See also Seedance API pricing compared (updated monthly), the evergreen guide this update will be folded into.


Seedance 2.5 Flash is a speed tier of Seedance 2.5 that BytePlus now lists on ModelArk under the model id dreamina-seedance-2-5-flash-261115. According to the ModelArk model list, it supports the same task types as Seedance 2.5: text to video, first frame, first and last frame, reference to video, video editing and video extension. The consumer app Dreamina also offers it, pitched there as faster previews at near original quality.
ModelArk bills video in output tokens. These are the list prices per million tokens for 480p and 720p, as shown on the BytePlus pricing page on 2026-11-16:
| Model | No video input | With video input | Highest resolution |
|---|---|---|---|
| Seedance 2.5 Flash | $7.0 | $4.3 | 720p |
| Seedance 2.5 | $10.70 | $6.40 | 4K |
| Seedance 2.0 | $7.0 | $4.3 | 4K |
| Seedance 2.0 Fast | $5.6 | $3.3 | 720p |
| Seedance 2.0 Mini | $3.5 | $2.1 | 720p |
So Flash costs the same per token as standard Seedance 2.0 and about 35% less than Seedance 2.5 at the same resolution. BytePlus's own worked example puts a 5 second, 16:9, 720p clip without video input at $0.76 on Flash against $1.156 on Seedance 2.5, and a 480p clip at $0.34 against $0.514. Requests that include video input have a minimum token charge, and BytePlus says you are charged only for videos that generate successfully.
Per the ModelArk docs as of 2026-11-16, Flash keeps the long clips of 2.5 but trims everything else:
We ran one prompt on both models on 2026-11-16: text to video, 5 seconds, 720p, a rain scene with a push-in. Both used about 108.9k tokens, which matches the BytePlus token formula for 5 seconds of 720p. Flash finished in 76 seconds against 139 for Seedance 2.5, and cost about $0.76 against $1.17 at list price.
Composition and prompt following held up on Flash. Detail was flatter. The character's hair stayed dry in the rain, and the push-in we asked for did not happen. That is one clip, so treat it as a hint, not a result.
Our read: Flash makes sense for blocking shots, testing prompts and drafts at 720p, where speed and price matter more than texture. For a final shot that needs fine detail, camera moves or 1080p, Seedance 2.5 is still the tier to use.
No, it is the same price per token. On BytePlus ModelArk both list at $7.0 per million tokens without video input and $4.3 with video input at 480p and 720p, as of 2026-11-16. Flash can make clips up to 30 seconds, while Seedance 2.0 stops at 15. Seedance 2.0 Fast and Mini are cheaper per token.
The API is not free. BytePlus ModelArk bills every successful generation by output tokens, and failed generations are not charged. Dreamina, ByteDance's consumer app, invites users to try Seedance 2.5 Flash there for free, but its plans and credits are separate from API pricing.
BytePlus's example puts a 5 second 720p clip without video input at $0.76, or about $0.15 per second, and a 5 second 480p clip at $0.34. Longer clips and clips with video input cost more.
On BytePlus ModelArk the model id is dreamina-seedance-2-5-flash-261115. Standard Seedance 2.5 is dreamina-seedance-2-5-260628.
This is a dated note. The prices here will be folded into our monthly Seedance API pricing comparison, which covers every provider. Dawnwood does not offer Flash today. Our current Seedance 2.5 rates are on the Seedance 2.5 model page.
Dawnwood's Auto mode turns a story told in a few lines into a vertical series. It writes the characters, the 10-episode outline and the scripts without spending credits, then films one episode at a time: you read the script and say go. Credits only go on art and filming, and each episode shows its cost before it starts.
Tell us your storyHow Auto mode works →