VideoAnonymized

Seedance 2.0 R2V

ByteDance's flagship multimodal reference-to-video model — accepts text, images, video, and audio inputs to generate cinematic 15-second clips with native audio.

Maker
ByteDance
Modality
Video + audio
Max duration
15 seconds
Max resolution
4k

Overview

What is Seedance 2.0 R2V

Seedance 2.0 R2V is ByteDance's flagship reference-to-video model, released in February 2026. It accepts up to 12 mixed inputs—including text, images, video clips, and audio—to generate coherent, audio-synchronized 4–15 second videos in a single pass, with cinematic camera control and character consistency.

Running it privately on Venice

On Venice, Seedance 2.0 R2V runs under an anonymized privacy tier—your prompts are not stored or profiled. This means you can generate video privately without building a personal history. The model’s full multimodal reference capabilities are accessible via API, letting you combine images, scripts, and audio without surveillance from Big Tech platforms.

AnonymizedNo prompt trainingTEE · hardware enclaveEnd-to-end encrypted

Assessment

Strengths and limitations

Strengths
  • Unmatched multimodal input support: combine text storyboards, reference images, audio tracks, and video clips in a single generation.
  • Native audio-video synchronization with dual-channel sound and ambient audio generation.
  • Strong character and motion consistency across multi-shot sequences, ideal for short-form social content.
  • Supports video editing, extension, and stitching workflows from a single endpoint.
  • Cinematic camera and lighting control inferred directly from multimodal prompts.
Limitations
  • No open weights: cannot be self-hosted or audited for bias or safety.
  • Aggressive content moderation outside controlled environments; person-bearing media not supported on public API.
  • Learning curve is steep for casual users: prompt engineering and reference structuring are essential.
  • Limited to 15-second outputs, which may not suit long-form storytelling needs.

Capabilities

What it supports

  • Image to video
  • Reference to video
  • Native audio generation

Specifications

Datasheet

Maker
ByteDance
Released
February 10, 2026
Modality
Reference-to-video (R2V), video editing, video extension, video stitching
Resolutions
4k, 1080p, 720p, 480p
Clip lengths
4s – 15s
Mode
image-to-video
Aspect ratios
21:9, 16:9, 4:3, 1:1, 3:4, 9:16
Audio
Yes
Prompt limit
10,000 chars
Privacy on Venice
Anonymized — prompts not stored
Available on Venice since
Mar 2026
License
Proprietary

API

Call it from your code

Venice exposes this model through the REST API. Queue a generation with the model id.

curl https://api.venice.ai/api/v1/video/queue \
  -H "Authorization: Bearer $VENICE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2-0-reference-to-video-basic",
    "prompt": "Aerial drone shot over a misty mountain valley at golden hour"
  }'

# Use the returned queue_id with https://api.venice.ai/api/v1/video/retrieve.
# Call /video/complete after downloading if needed.

Pricing

What it costs on Venice

Pay per clip on Venice — price scales with resolution and duration (4s–15s), from $0.76.

4k · 4s
$3.89
Per clip
1080p · 4s
$1.87
Per clip
720p · 4s
$0.76
Per clip

New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.

Alternatives

How it compares

ModelMax resolutionStrongest atOpen weightsPrice (Venice)
Seedance 2.0 R2V4kMultimodal reference controlNofrom $0.76
Kling O3 Pro4kCinematic realismNofrom $0.46
Wan 2.7 Enhanced1080pSpeed & volumeYesfrom $0.68
Vidu Q32kNarrative coherenceNofrom $0.27

Best for combining images, audio, and text into cinematic clips with high character consistency.

Use cases

What it is good for

  1. 01Short-form social videos for TikTok, Reels, and YouTube Shorts with consistent characters and branded audio.
  2. 02Pre-visualization and storyboarding using text scripts and reference images.
  3. 03Dynamic ad creatives combining product shots, voiceovers, and background music.
  4. 04Music video prototyping with synced visuals and audio stems.
  5. 05Multimodal editing workflows where video clips are extended or restyled using reference assets.

Prompting

Getting better results

Structure your prompt to clearly separate text instructions from file references — use labels like 'REFERENCE_IMAGE_1', 'AUDIO_TRACK'.

Include timing cues (e.g., 'at 0:03, character turns') to improve camera and action alignment.

Use high-quality reference images for characters and scenes to maximize consistency.

For audio, provide a clear description or upload a reference track to guide mood and pacing.

Version history

Seedance 1.0
2024

Initial release with basic image-to-video capabilities.

Seedance 2.0
2026-02

Major upgrade with unified multimodal architecture and R2V support.

Seedance 2.5
2026-07

Next iteration with improved motion coherence and longer clip support.

FAQ

Frequently asked questions

Seedance 2.0 R2V is ByteDance's reference-to-video model that generates 4–15 second clips from a combination of text, images, video, and audio inputs. It supports editing, extension, and stitching workflows and produces cinematic output with native audio.

On Venice, pricing starts at $0.76 per clip for 720p at 4 seconds, scaling up to $3.89 for 4k at the same duration. Prices vary based on resolution, length, and aspect ratio.

No. Seedance 2.0 R2V is a proprietary model developed by ByteDance. It is not open source or freely available for self-hosting.

Yes. It supports native audio-video joint generation, including ambient sound and synced audio tracks. You can include audio files as references in your prompt.

It supports 480p, 720p, 1080p, and 4k resolutions, with flexible aspect ratios including 16:9, 9:16, and cinematic 21:9.

Yes. It supports video editing, extension, and stitching workflows — you can modify specific segments while preserving motion, style, and character consistency.

Seedance 2.0 R2V excels in multimodal input flexibility and character consistency; Kling O3 Pro leads in cinematic realism and physics accuracy. Choose Seedance for reference-heavy workflows, Kling for pure visual fidelity.

Run Seedance 2.0 R2V privately

No prompt logging. No data used for training.