SeeDance 2.5 Flash API

Updated 13 Nov 2026

What is SeeDance 2.5 Flash?

SeeDance 2.5 Flash is the faster, lower-cost variant of ByteDance's SeeDance 2.5 video model. It is live on UnificAlly as bytedance/seedance-2.5-flash and generates 4 to 30 second videos at 480p or 720p, with synchronized audio on by default.

The request is the same as SeeDance 2.5's. Three workflows share one model ID: text-to-video, first-and-last-frame generation, and multimodal reference generation with images, videos, and audio in one call. Code written for SeeDance 2.5 runs on Flash after you change the model string, as long as the request stays inside Flash's limits.

Those limits are the trade. Flash stops at 720p and takes 20 references per request, where SeeDance 2.5 reaches 1080p and takes 50. In return, each second of video costs less and comes back sooner.

Key features of SeeDance 2.5 Flash

What SeeDance 2.5 Flash is best for

• Prompt and reference tests

  Try several versions of a scene at the lower rate, then run the winning request on SeeDance 2.5 when it needs 1080p.

• Social video at 720p

  Vertical 9:16 output with sound is ready for feeds, stories, and ads that never need more than 720p.

• High-volume generation

  A lower price per second makes large batches of product, ad, or catalog videos cheaper to produce.

• Longer videos on a budget

  The 30-second ceiling is the same as SeeDance 2.5, so a full short story fits in one request at Flash rates.

• Reference-led scenes

  Combine product images, a motion reference, and a voice or music cue in one request with numbered references.

SeeDance 2.5 Flash use cases

Build an ad pipeline that drafts ten versions of a 15-second product video at 480p, lets a person choose one, and sends only that request to SeeDance 2.5 for the 1080p render. Generate vertical 720p videos with dialogue for social channels straight from a script. Give an e-commerce catalog a short video per product by passing the product photos as [Image1] and [Image2] with a shared camera move in [Video1]. For an ad with a fixed open and close, pin the first frame and the final call-to-action frame and let the model fill in the motion. Audio-led work can start from a voice or rhythm reference with no image at all.

SeeDance 2.5 Flash limitations

Output is 480p or 720p only. Flash has no 1080p option and no draft mode, so a Flash video cannot be re-rendered at a higher resolution from its task ID. The reference budget is 10 images, 5 videos, and 5 audio files, 20 in total, which is less than half of SeeDance 2.5's. Reference videos need at least 409,600 pixels per frame and can total 30 seconds, and every one of those seconds is billed like a second of output. First-and-last-frame mode needs both images and cannot be mixed with references. Arena's text-to-video board has no SeeDance 2.5 Flash row as of November 13, 2026, so there is no independent rank or Elo score for this variant yet.

When to use SeeDance 2.5 Flash

Use SeeDance 2.5 Flash when 720p is enough and cost per second or turnaround matters more than the last step of resolution: drafts, social video, and batch jobs. Use SeeDance 2.5 when the video needs 1080p, more than 20 references, or draft mode. Because both models take the same request, a common setup is to test on Flash and render the final version on SeeDance 2.5.

What SeeDance 2.5 Flash can do

Both videos below come from one prompt, sent with the same settings to each model: 720p, 8 seconds, 16:9, seed 20261113, audio on. The exact prompt:

Night street-food market in light rain. Handheld camera starts wide on a crowded alley under a glowing neon sign that reads "NOODLES 24H", then pushes in to a close-up of a woman chef at her stall. She tosses noodles in a wok and a burst of flame flares up, she pours hot broth from a ladle into a bowl as steam rises, sprinkles chopped scallions over it with her fingers, then looks into the camera, smiles and says "Order up!". People with umbrellas walk past behind her, neon reflects on the wet ground. Realistic, cinematic, natural sound of sizzling, rain and crowd.

SeeDance 2.5 Flash API examples

Call SeeDance 2.5 Flash from any language by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/video/bytedance/seedance-2.5-flash.

curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "bytedance/seedance-2.5-flash",
    "input": {
      "prompt": "A red kite dancing over a green hill on a sunny afternoon, wind rustling through the grass",
      "aspect_ratio": "16:9",
      "duration": 8,
      "resolution": "720p",
      "generate_audio": true
    }
  }'

Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished result.