Video · ByteDance

Seedance 2.5

ByteDance's July 2026 model, which makes up to 30 s in one pass with sound. Pick it for a long continuous scene.

Anonymized, some versions private Moderate cost
Latest versionSeedance 2.5July 31, 2026
Artificial Analysis Text to Video Arena with audio1,142 Elo#2 of 18 on Artificial Analysis Text to Video Arena with audio
Price$0.13 per second of video$0.648 a 5 s clip
Key specClips up to 30 s, with sound

What sets it apart

Seedance 2.5 is ByteDance's long-clip model: 30 s in a single run, sound generated with the picture, and a reference budget of up to 30 images, 10 videos and 10 audio clips. It can also be extended turn by turn.

  • 30 s per clip, extendable
  • Up to 30 images, 10 videos, 10 audio
  • Sound generated with the picture

Recommended for: one long continuous shot.

Strengths and limits

Where it shines

  • Up to 30 s in a single generation, no stitching
  • Sound generated with the picture
  • The only Seedance with a private (zero retention) version

Think twice when

  • Among the dearest per second; the private US version costs about twice as much
  • Public versions refuse photos with a recognizable person
  • No reference clips, only images and audio
  • 4 to 30 s, native sound
  • Up to 1080p
  • Text, image or references to video
  • Public versions anonymized, US versions private

Use cases

Whole scene in one take

A 25 second scene in a night market: the camera follows a vendor frying noodles, a customer arrives and orders in English, steam rises, crowd noise and sizzling throughout.

30 s per clip removes stitching for a full scene.

Cast from reference images

Use my three character reference images and the audio clip of the narrator's voice. The trio walks through a library while the narrator speaks.

It accepts many references at once for several characters.

Music-driven clip

A 15 second clip of a skateboarder in an empty parking lot, cuts on each beat of my reference track, warm sunset light.

Audio references can steer timing and tempo shifts.

Against its rivals

  • Wan 3.0Wan 3.0 also reaches 30 s and takes documents as input.#3 of 22 on Artificial Analysis Image to Video Arena (1,164 Elo)
  • Kling V3Kling V3 is shorter (15 s) but adds motion control and a 4K tier.#12 of 22 on Artificial Analysis Image to Video Arena (1,055 Elo)
  • MiniMax H3 (Hailuo)MiniMax H3 has open weights and 2K output at 15 s.#1 of 22 on Artificial Analysis Image to Video Arena (1,195 Elo)
Compare side by side ↗

Version history

Each version, dated, with what it changed over the one before.

  1. Seedance 2.0Earlier, not in Sub Rosa
    • Up to 15 s clips with native audio and multi-shot scenes from text, image, audio and video.
    Source : en.wikipedia.org
  2. Seedance 2.5Latest
    • Clip length doubles to 30 s in a single run, with multi-turn extension beyond.
    • Reference budget up to 30 images, 10 videos and 10 audio clips per request.
    • Timestamp-specific editing that keeps the visual style.
    Source : technode.com

Benchmarks

How to read these scores: what a benchmark tells you.

Artificial Analysis Text to Video Arena with audio: #2 of 18 models measured in the catalog. The duel for clips with sound: picture, voices, effects and how well they match. An Elo rating: higher is preferred more often. Only models that make their own sound take part.

Artificial Analysis Text to Video Arena with audioElo rating · Artificial Analysis, October 4, 2026
  1. 1. Wan 3.0Wan 3.01,156
  2. 2. Seedance 2.5Seedance 2.51,142
  3. 3. MiniMax H3MiniMax H3 (Hailuo)1,138
  4. 4. MiniMax H3 MaxMiniMax H3 (Hailuo)1,134
  5. 5. Flux 3FLUX 3 video1,127
  6. 6. Seedance 2.0Seedance 2.0 and 1.51,119
  7. 7. Gemini Omni Flash 1.1Gemini Omni Flash1,113
  8. 8. Grok Imagine 1.5Grok Imagine video1,044
  9. 9. HappyHorse 1.1HappyHorse1,043
  10. 10. Wan 2.7Wan 2.x1,029
See the numbers
ModelElo
1. Wan 3.01,156
2. Seedance 2.51,142
3. MiniMax H31,138
4. MiniMax H3 Max1,134
5. Flux 31,127
6. Seedance 2.01,119
7. Gemini Omni Flash 1.11,113
8. Grok Imagine 1.51,044
9. HappyHorse 1.11,043
10. Wan 2.71,029

Artificial Analysis Text to Video Arena: #4 of 5 models measured in the catalog. People watch two silent clips made from the same description, without knowing the models, and pick the better one. An Elo rating: higher is preferred more often. 100 points apart means the higher-rated model wins about 64% of duels.

Artificial Analysis Text to Video ArenaElo rating · Artificial Analysis, October 4, 2026
  1. 1. Wan 3.0Wan 3.01,129
  2. 2. MiniMax H3MiniMax H3 (Hailuo)1,118
  3. 3. MiniMax H3 MaxMiniMax H3 (Hailuo)1,115
  4. 4. Seedance 2.5Seedance 2.51,102
  5. 5. Flux 3FLUX 3 video1,094
See the numbers
ModelElo
1. Wan 3.01,129
2. MiniMax H31,118
3. MiniMax H3 Max1,115
4. Seedance 2.51,102
5. Flux 31,094

All 6 models in this family

ModelUsePrivacyPrice
Seedance 2.5seedance-2-5-text-to-video-basicText to videoAnonymized$0.13 per second of video$0.648 a 5 s clip
Seedance 2.5seedance-2-5-image-to-video-basicImage to videoAnonymized$0.13 per second of video$0.648 a 5 s clip
Seedance 2.5 R2Vseedance-2-5-reference-to-video-basicReferences to videoAnonymized$0.13 per second of video$0.648 a 5 s clip
Seedance 2.5 USseedance-2-5-us-text-to-video-privateText to videoPrivate$0.281 per second of video$1.403 a 5 s clip
Seedance 2.5 USseedance-2-5-us-image-to-video-privateImage to videoPrivate$0.281 per second of video$1.403 a 5 s clip
Seedance 2.5 US R2Vseedance-2-5-us-reference-to-video-privateReferences to videoPrivate$0.274 per second of video$1.372 a 5 s clip

Specifications

SpecificationSeedance 2.5
ReleasedJuly 31, 2026
WeightsClosed
Acceptstext, images, video, audio
Producesvideo, audio
Clip length4 to 30 s, with sound
Resolutions1080p, 720p, 480p
Price$0.13 per second of video
PrivacyAnonymized

Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.