On November 12, ByteDance's BytePlus and Volcano Engine jointly launched the Dreamina Seedance 2.5 Flash video generation model. ByteDance calls it the closest model yet to flagship Seedance 2.5 in quality, says it generates nearly twice as fast, and says it beats the previous flagship Seedance 2.0 overall. BytePlus prices it at $0.070 per second at 480p and $0.152 per second at 720p, while Volcano Engine in China charges about 0.99 yuan per second at 720p; both are roughly 34% below Seedance 2.5. A single clip runs up to 30 seconds with up to 10 images, 5 videos and 5 audio clips as references. The launch covers only 480p and 720p, without Draft mode. It is rolling out on Dreamina and partner platforms now, and the ModelArk and Volcano Ark APIs open on November 15 at 14:00 (UTC+8).

Key Takeaways

  • ✓Speed and quality: about 2x the generation speed of Seedance 2.5; ByteDance says quality is the closest yet to the flagship and better than the previous flagship Seedance 2.0 overall (vendor claim, no third-party benchmark yet)
  • ✓Global pricing: $0.070/sec at 480p and $0.152/sec at 720p, versus $0.103 / $0.231 for flagship Seedance 2.5, so 720p is about 34% cheaper (BytePlus comparison chart)
  • ✓China pricing: about 0.99 yuan/sec for 720p text-to-video and image-to-video on Volcano Engine, also about 34% below Seedance 2.5
  • ✓Specs: clips up to 30 seconds, up to 10 images + 5 videos + 5 audio clips as references (flagship allows 30+10+10); launch resolutions are 480p/720p only, no Draft mode
  • ✓Availability: rolling out on Dreamina, LibTV, Updream, Creao AI and TapNow now; BytePlus ModelArk and Volcano Ark APIs open November 15, 14:00 (UTC+8)
ByteDance Launches Seedance 2.5 Flash Video Model: Close to Flagship Seedance 2.5 Quality at Nearly 2x Speed, $0.152 per Second at 720p (About 34% Cheaper), API on November 15
🖼️Official Media
Click to view high-res
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

🔬

In-Depth Technical Analysis

Background

Video generation is moving from one-off showcase clips to high-volume ads, short dramas and e-commerce content, where cost and wait time matter more than peak quality. Flagship Seedance 2.5 already offers 1080p and a 4K preview, but at $0.231 per second for 720p it is expensive for iterating on shots or producing many ad variants. Seedance 2.5 Flash targets that high-frequency, high-volume work.

Architecture highlights

ByteDance has not disclosed the architecture or parameter count. It describes Flash as keeping most of the flagship's creative capabilities in exchange for faster generation and a lower price. It inherits the 2.5 family's multimodal reference-to-video and video editing features with clips up to 30 seconds, but reference limits drop from the flagship's 30 images/10 videos/10 audio clips to 10/5/5, launch resolutions are 480p and 720p only, and the flagship's Draft mode (480p draft first, then a final render) is not supported.

Benchmarks

No benchmark scores have been published, by ByteDance or by third-party leaderboards. The confirmed claims are ByteDance's own: about 2x the generation speed of Seedance 2.5, better overall results than Seedance 2.0, and strong instruction following, physical motion and audio. On price, 720p drops from $0.231 to $0.152 per second (about 34%) and 480p from $0.103 to $0.070 (about 32%).

Getting started

From November 15 at 14:00 (UTC+8), call it through the video generation task API on BytePlus ModelArk (global) or Volcano Ark (China) as dreamina-seedance-2-5-flash. A practical workflow is to generate 480p/720p variants and pick shots with Flash, then render the final 1080p or 4K version with flagship Seedance 2.5. Until the API opens, you can try it on Dreamina and partner platforms.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

Up NextMBZUAI's IFM Open-Sources PAN 2 World Model: A Multimodal LLM Plans Who Does What and When, a 29B Diffusion Model Renders Up to 20 Seconds of Simulation, Top Open-Weight Physics-IQ Score of 49.38, Apache 2.0
Read Now→