MakeJam
Download

ByteDance

Seedance 2.5

  • Video
VideoByteDance
Open in Studio

Description

ByteDance Seedance 2.5 on WaveSpeed.

Seedance. Video generation from ByteDance

Tags

seedancebytedanceaudiophysicslong-takemcp

Capabilities

Vendor
ByteDance
Family
Seedance
Type
Video
Model ID
Tasks
fal.ai: Text to video, Image to video, Reference to video · Higgsfield: Text to video, Image to video, Reference to video · KREA: Text to video, Image to video, Reference to video · Pika: Text to video, Image to video, Reference to video · Venice: Text to video, Image to video, Reference to video · WaveSpeed: Text to video, Image to video, Video to video
Draft
fal.ai: Preview switch. · Higgsfield: Preview and finishing a draft are separate, and cannot be combined. Id field: draft_job_id. · KREA: Preview switch.
Extend
Higgsfield: forward, backward · WaveSpeed: forward
Seed
KREA: Can be set.

Inputs

Prompt
Optional
Image
Optional
Video
Optional
Audio
Optional
End frame
Supported
Max images
30
Reference images
Up to 30 reference images.
Reference videos
Up to 10 reference videos.
Reference audio
Up to 10 reference audio clips.
Reference total
Up to 50 references in total.
Elements
Up to 30 elements, 30 images each, 1 video each, 1 audio clip each.
Source video
Accepted.
Source audio
Accepted.

Outputs

Duration
fal.ai: auto, 4–30 seconds · Higgsfield: 4–30 seconds · KREA: 4–30 seconds (default 5) · Pika: 4–30 seconds · Venice: 4–30 seconds · WaveSpeed: 4–30 seconds (default 5)
Resolution
fal.ai: 480p, 720p, 1080p · Higgsfield: 480p, 720p, 1080p · KREA: 480p, 720p, 1080p · Pika: 480p, 720p, 1080p · Venice: 480p, 720p · WaveSpeed: 480p, 720p, 1080p, 4k
Aspect ratios
fal.ai: auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 · Higgsfield: auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 · KREA: 16:9, 4:3, 1:1, 3:4, 9:16, 21:9 · Pika: 1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive · Venice: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 · WaveSpeed: 16:9, 9:16, 4:3, 3:4, 1:1, 21:9
Audio
fal.ai: Generated, and the catalog records a switch. · Higgsfield: Generated, and the catalog records a switch. · KREA: Generated. The catalog records no switch. · Pika: Generated. The catalog records no switch. · Venice: Generated, and the catalog records a switch. · WaveSpeed: Generated, and the catalog records a switch.
Catalog description
$0.2205/s

Token pricing, and audio does not change it: about $0.2205/s at 480p, $0.4730/s at 720p, $1.164/s at 1080p ($0.0214 per 1,000 tokens at 480p and 720p, $0.0234 at 1080p).

Full prices on modelpricetrack.com

Where to run it

Tasks and limits are per catalog row. A provider may offer only some of them.

fal.aiAPI

Video

End frame
Supported.

HiggsfieldMCP

Video

KREAAPI

Video

End frame
Supported.

PikaAPI

Video

VeniceAPI

Video

WaveSpeedAPI

Video

End frame
Supported.

Prices

Only sentences in the catalog that state an amount and a unit.

  • Token pricing, and audio does not change it: about $0.2205/s at 480p, $0.4730/s at 720p, $1.164/s at 1080p ($0.0214 per 1,000 tokens at 480p and 720p, $0.0234 at 1080p).

    Source: @diune-sas/ai-providers catalog description · fal.ai

  • Video references multiply the price by 0.6 and bill both input and output duration (about $0.2838/s at 720p with video).

    Source: @diune-sas/ai-providers catalog description · fal.ai

  • Per second, without / with a video reference: 480p $0.1027 / $0.0614, 720p $0.2311 / $0.1382, 1080p $0.5686 / $0.3402.

    Source: @diune-sas/ai-providers catalog description · KREA

  • Text-to-video and image-to-video, per second: $0.18 at 480p, $0.36 at 720p, $0.90 at 1080p, $1.80 at 4k.

    Source: @diune-sas/ai-providers catalog description · WaveSpeed

  • With reference video on text-to-video the rate is $0.11 / $0.22 / $0.55 / $1.10 per second of normalized input plus output.

    Source: @diune-sas/ai-providers catalog description · WaveSpeed

Use cases

  • A clip from a prompt

    A slow push-in through a rainy alley, neon signs reflected in the pavement

  • Motion from a still

    The framed subject turns slightly as the light shifts

Tips

  • Every catalog row allows a clip from 4 to 30 seconds, lists 480p and 720p, accepts up to 30 reference images, and generates audio.