Seedance 1.0 guide: release, capabilities and current status
Quick answer
Seedance 1.0 is ByteDance Seed’s June 2025 visual video model for text-to-video, image-to-video and native multi-shot generation. It remains useful for understanding older results, but it is no longer the current Seedance release: Seedance 1.5 Pro added audio, Seedance 2.0 added multimodal references and editing, and Seedance 2.5 added up to 30-second generation and multi-round extension. For a new project, evaluate 2.5 first unless your provider or workflow specifically requires 1.0. Verified September 13, 2026 from ByteDance Seed sources.
Seedance 1.0 does not have one universal free plan or price. ByteDance developed the model, while each consumer app, API or third-party host controls access, limits, billing and the settings it exposes. Confirm the exact provider and version before following any “free Seedance” tutorial.
Seedance versions at a glance
Version | Official release | Documented change | Use now |
|---|---|---|---|
June 11, 2025 | Text-to-video, image-to-video and native multi-shot visual generation | Use for historical 1.0 evaluation or a host that explicitly exposes it | |
December 2025 | Added joint audio-video generation and multilingual or dialect speech | Use only when its particular host and controls fit the job | |
February 12, 2026 | Added unified text, image, audio and video reference and editing inputs | Retain for existing tests; compare with 2.5 for new work | |
July 31, 2026 | Added up to 30-second generation, multi-round extension and stronger reference editing | Current starting point for a new Seedance evaluation |
Try the current image-to-video workflow
Start with an approved reference image and a short shot brief, then select a currently available model and verify the complete output before scaling the workflow.
Try Image-to-VideoWhat Seedance 1.0 introduced
ByteDance’s Seedance 1.0 model page describes a model that generates video from text and images, supports native multi-shot storytelling and produces 1080p output. The model’s purpose was to improve prompt following, motion stability, visual detail and continuity across connected shots.
The official technical report documents a unified model for text-to-video and image-to-video, interleaved multimodal positional encoding, multi-stage distillation and video-specific post-training. Use that report instead of repeating unsourced training-data totals or architecture labels from third-party summaries.
What the official speed figure means
ByteDance’s release note reports 41.4 seconds to generate one five-second 1080p video on an NVIDIA L20 in its own test environment. That is a provider benchmark, not a promise that a web app will return every clip in 41.4 seconds.
End-to-end time also includes queueing, upload, preprocessing, safety review, host configuration and download. If speed affects delivery, measure from request submission to a usable file on the exact provider and account you plan to use.
Seedance 1.0 is visual-only
Seedance 1.0 generates the visual clip. Native synchronized audio arrived in Seedance 1.5 Pro. Do not attribute dialogue, sound effects or music generation to a 1.0 result unless a separate tool or hosting platform added that layer.
For a 1.0 workflow, prepare the visual shot first, then add licensed music, voiceover or sound effects in editing. Keep the audio source and rights record with the final project.
How to use Seedance 1.0 accurately today
Confirm the version. Look for an explicit Seedance 1.0 model label; a page saying only “Seedance” may have moved to a newer version.
Choose the input. Use text-to-video for a written scene or image-to-video when an approved image should anchor the subject or first frame.
Write a shot brief. State subject, action, setting, camera, lighting, duration and aspect ratio. Ask for multiple shots only when the selected host exposes that behavior.
Generate a small test. Inspect the entire output for subject drift, motion errors, invented objects, transitions and readable details.
Record the provider. Save host, exact model, settings, prompt, source asset, charge and output. Provider controls can materially change a result.
A practical multi-shot prompt
Ten-second cinematic sequence of an unbranded red hiking backpack on a wooden bench at sunrise. Shot 1: wide establishing view. Shot 2: medium push toward the backpack. Shot 3: close detail of the zipper. Keep the same backpack shape, color and fabric across all shots; natural camera movement; 16:9.
This is an illustrative prompt, not a retained Magic Hour benchmark result. Use an authorized reference image when exact product appearance matters, and add on-screen text after generation so it remains editable and readable.
When to use a newer Seedance version
Seedance 2.0 is a better comparison when the job needs mixed text, image, audio or video references, editing or native audio-video generation. Seedance 2.5 is the current starting point when you need longer output, multiple rounds of extension or stronger reference-based editing.
Do not relabel a historical 1.0 output as 2.5. Keep the original version attached to any example, benchmark or cost figure. When updating a workflow, run the same permitted brief through the newer version and compare accepted outputs, total spend and correction time.
Access and pricing checks
ByteDance’s current model index still lists Seedance 1.0 alongside later releases, but a model listing does not establish free use in a particular app. Before subscribing or integrating, verify the host, region, model ID, input modes, duration, resolution, audio, commercial terms, queue, watermark and current quote.
Avoid converting a subscription allowance into a per-clip claim unless the credit cost for the exact model and settings is documented. Count rejected generations when estimating cost per accepted video.
Frequently asked questions
There is no universal Seedance 1.0 free plan. Free access, signup requirements, credits, watermarks and export limits depend on the provider and can change.
ByteDance published the Seedance 1.0 technical report and release note on June 11, 2025.
No native audio capability is documented for 1.0. Seedance 1.5 Pro introduced joint audio-video generation.
Yes. ByteDance documents native multi-shot generation and shows ten-second examples with two to three transitions. The exact duration and controls exposed to you depend on the host.
Start with 2.5 for a new evaluation. Use 1.0 when you must reproduce a historical result, compare versions or use a provider that explicitly requires it.
Related guides
For current release dates, use the AI video model release tracker. For current 2.0 access and costs, see Seedance 2.0 pricing. For a broader platform decision, compare the best AI video generators.
