As of: 2026-10-25 23:49 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/m/seedance-2-5/

# Dreamina Seedance 2.5

ByteDance Seed, Seedance family. Status: current. Type: video generation. Released: 2026-07-31.

## How to call it

| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| BytePlus ModelArk | `dreamina-seedance-2-5-260628` | https://ark.ap-southeast.bytepluses.com/api/v3/ | https://docs.byteplus.com/en/docs/modelark/seedance-2-5 |
| OpenRouter | `bytedance/seedance-2.5` |  |  |
| Web app (Jimeng / Doubao Pro, China) |  |  | https://seed.bytedance.com/en/seedance2_5 |
| Web app (Dreamina, CapCut) |  |  | https://dreamina.capcut.com/seedance/seedance-2-5 |

## Pricing

- Per 1m tokens 480p 720p: $10.70
- Per 1m tokens 1080p: $11.70
- Per 1m tokens with video input 480p 720p: $6.40
- Per 1m tokens with video input 1080p: $7
- Unit as published: USD per 1M video tokens on BytePlus ModelArk (online inference; 4K not supported). Tokens = (input + output video seconds) x width x height x fps / 1024. BytePlus examples, 5 s 16:9 without video input: 480p 0.514 (0.103/s), 720p 1.156 (0.231/s), 1080p 2.843 (0.569/s). Draft mode is billed as 480p
- Source: https://docs.byteplus.com/en/docs/modelark/model-pricing
- Last checked: 2026-10-25

## Specs

- Input: text, image, video, audio
- Output: video, audio
- Context window: not published
- Max output: not published
- Open weights: no
- Licence: proprietary

## Where it stands

- Newer video generation models in the registry since its release: 7. The 5 newest:
  - FLUX 3 Max (2026-10-19) https://postcutoff.com/m/flux-3-max/
  - Kling 4.0 (and 4.0 Flash) (2026-10-15) https://postcutoff.com/m/kling-4-0/
  - Tavus Griffin (Human Interaction Model) (2026-10-01) https://postcutoff.com/m/tavus-griffin/
  - HeyGen Video 1 (2026-09-30) https://postcutoff.com/m/heygen-video-1/
  - Utopai X (2026-09-30) https://postcutoff.com/m/utopai-x/
- Before it from ByteDance Seed: Dreamina Seedance 2.0 (2026-02) https://postcutoff.com/m/seedance-2-0/
- Closed weights; access is hosted only.

## What stands out

- 30-second single-shot generation with extensions: Up to 30 s of joint audio-video per generation (15 s in Seedance 2.0), extendable twice with consistent characters, settings and pacing. Source: https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5
- 50 multimodal references in one pass: Up to 30 images, 10 video clips and 10 audio clips as references in a single request. Source: https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5
- Timestamp-level audio/video editing: Targeted edits of audio and video at given timestamps, plus green-screen, white-model (untextured previs) control and camera-perspective editing. Source: https://seed.bytedance.com/en/seedance2_5
- Draft-then-final workflow (found after launch): API can render a cheap 480p Draft first and then generate the final video from the Draft task ID (final currently 1080p only; Draft IDs valid 7 days). Source: https://docs.byteplus.com/en/docs/modelark/seedance-2-5

## Notes

Artificial Analysis text-to-video leaderboard (read 2026-10-25): #5 (rank range 2-8), Elo 1158; AA price 34.12 USD per minute of 1080p (matches BytePlus 0.569/s). Launched 2026-07-31 in Jimeng and Doubao Pro with API 'coming soon'; the ModelArk id dreamina-seedance-2-5-260628 is now in BytePlus docs and pricing (API go-live date not verified; OpenRouter lists it from 2026-08-07). Third-party pages claim native 4K, but BytePlus pricing marks 4K 'Not supported' and the Seed pages give no resolution. Dreamina's 180 s 'long video' beta not verified on Seed pages. Launch entry: 2026-07-31-seedance-2-5. Oct 23, 2026: Volcano Engine opened 'Seedance 2.5 4K' (4K 10-bit, 480p draft previews) to selected enterprise customers only (TechNode citing IT Home); no public 4K API id or price seen.

ByteDance Seed's flagship audio-video model (successor to Seedance 2.0): 30-second single shots, two extensions, up to 50 references, timestamp-level editing. Asynchronous ModelArk content-generation task API; activate the model in the ModelArk console first.

Artificial Analysis text-to-video (with audio) leaderboard, read 2026-10-25: rank 5 (CI range 2-8), Elo 1158 (https://artificialanalysis.ai/video/leaderboard/text-to-video).

Sources: https://seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-introducing-seedance-2-5 · https://seed.bytedance.com/en/seedance2_5 · https://docs.byteplus.com/en/docs/modelark/seedance-2-5 · https://docs.byteplus.com/en/docs/modelark/model-pricing · https://technode.com/2026/07/31/bytedance-launches-seedance-2-5-video-generation-model/

Model page: https://postcutoff.com/m/seedance-2-5/