Newshot
Account

ByteDance's Volcano Engine launches Seedance 2.5 flash video model: twice the generation speed at a lower price

AI summary

ByteDance's Volcano Engine officially launched Seedance 2.5 flash, which it says is the model closest in quality to Seedance 2.5, with generation twice as fast and a lower price. For text-to-video and image-to-video at 720P, Seedance 2.5 flash costs about 0.99 yuan per second, roughly 34% less than Seedance 2.5. It supports clips of up to 30 seconds, accepts up to 10 images, 5 videos and 5 audio clips per input, and offers 480P and 720P resolutions; it will roll out to platforms including Jimeng, LibTV, Updream, Creao AI and TapNow, with the API going live on Volcano Ark on Sunday, November 15.

Translation(model translation, not the source's own text)

ByteDance's Volcano Engine Releases Seedance 2.5 flash Video Model: Twice the Generation Speed, Lower Price

IT Home, November 12 — ByteDance's Volcano Engine officially launched Seedance 2.5 flash today.

According to the company, this is the model whose results come closest to Seedance 2.5, with generation speed doubled and a lower price, offering a more cost-effective option for batch content production. Taking text-to-video and image-to-video at 720P resolution as an example, Seedance 2.5 flash costs about 0.99 yuan per second to generate, roughly 34% less than Seedance 2.5.

Seedance 2.5 flash supports generating videos up to 30 seconds long per clip, accepts up to 10 images, 5 videos and 5 audio clips per input, and offers two resolution tiers: 480P and 720P.

Starting today, Seedance 2.5 flash will roll out on platforms including Jimeng, LibTV, Updream, Creao AI and TapNow; the API will go live on Volcano Ark this Sunday (November 15).

According to the introduction, Seedance 2.5 flash outperforms the previous-generation flagship model Seedance 2.0 overall and is currently the model whose results come closest to Seedance 2.5. It maintains a high standard in instruction following, physical motion and audio performance, with natural, textured visuals; its excellent video editing capabilities let users flexibly adjust content expression and produce different creative versions.

To suit different usage scenarios, the Seedance 2.5 family currently offers two tiers: the flagship model and the flash model:

  • The flagship model Seedance 2.5 represents the strongest intelligence available today, meeting higher demands for audiovisual quality and creative expression; combined with the Draft mode workflow of "pick a sample first, then generate the final cut," it can also reduce high-definition generation costs during repeated trial and error, serving 1080P and 4K high-resolution production.
  • Seedance 2.5 flash delivers results close to the flagship, with faster generation and lower cost, offering a cost-effective option for high-frequency creation and batch content production.

IT Home attaches the differences between the two models below:

Related Articles

  • Doubao PR head Liu Xing responds to "Hengdian and ByteDance cooperation to feed the portraits and voices of 150,000 extras to a large model": the rumor is seriously inaccurate
  • ByteDance's Volcano Engine releases Seedance 2.5 4K model, aimed at film-grade visual generation
  • ByteDance's Volcano Engine launches Seedance 2.5 4K enterprise trial, with the first batch supplied to 20 studios on a targeted basis
  • ByteDance's Volcano Engine launches the Seedance premium AI drama incentive program, with up to 200,000 yuan per work
  • ByteDance's Volcano Engine releases the Seedance film and television cooperation program: offering Token subsidies, with up to one million in incentives per project
  • Seedance 2.5 launches Draft mode: first validate ideas with 480P, then generate the final version in 1080P once satisfied

The translation is machine-generated and individual terms may be off. Nothing is rewritten or cut; the original is authoritative.

Published by IT之家
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗