Seedance 2.0 Mini AI Video Generator

Seedance 2.0 Mini has its own 480p and 720p pricing, 4-to-15-second output range, and smaller reference limits. It is suited to compact multimodal drafts, while Seedance 2.5 adds 1080p, longer output, and a much larger reference capacity.

Text uses a prompt of 3 to 20,000 characters with no uploads. Frames uses the same prompt range plus exactly one starting image or two images for a start-and-end pair, with no video or audio. Reference requires at least one uploaded item and can combine up to nine images, three videos, and three audio files.

Images may be JPEG/PNG/WebP/GIF up to 30 MB each. Reference videos must be MP4/MOV/MKV up to 50 MB each and 2 to 15 seconds long, with no more than 15 video seconds in total. MP3 or WAV audio is limited to 15 MB and 2 to 15 seconds per file, with no more than 15 audio seconds in total.

Images and reference videos must each have width and height from 300 to 6,000 pixels and an aspect ratio from 0.4 to 2.5. Reference video must also contain 409,600 to 927,408 pixels in total and run at 24 to 60 frames per second.

Without a reference video, the rate is 3.8 credits per output second at 480p or 8.2 at 720p. With video input, the rate is 2.4 at 480p or 5 at 720p and multiplies the verified input-video seconds plus requested output seconds. Fractional rates stay intact and Rivya rounds the combined total up once.

Web search is accepted only in the text scene. A frames or reference request with web search enabled is rejected, so switch it off before adding guide media.

Rivya reserves the full formula before generation. When valid actual usage at or below the reservation is reported, Rivya settles to it. Missing, invalid, or higher-than-reserved usage remains in reconciliation instead of triggering a hidden extra charge.

Choose a text-only, frame-guided, or multimodal reference scene; write a prompt of at least three characters; then set 480p or 720p, aspect ratio, and a whole-number duration from 4 to 15 seconds.

No. The frames scene accepts one or two images and rejects video or audio. Use the reference scene instead when images, videos, or audio need to guide the same clip.

The current Rivya limit is 15 files in total: up to nine images, three videos, and three audio files. Video and audio also have separate 15-second total-duration limits.

Yes. A request with video input uses the lower with-video rate but bills that rate across the verified input-video seconds plus the requested output seconds. Images and audio alone keep the no-video formula.

Stay with Mini for 480p or 720p drafts up to 15 seconds and smaller reference sets. Consider Seedance 2.5 when the task needs 1080p, up to 30 seconds, automatic duration, MOV output, or a larger multimodal bundle.