Technical specifications · ByteDance Seedance 2.0

Seedance 2.0 Specs: Resolution, Duration, Audio & API Pricing

Everything you need to know about Seedance 2.0 output quality in one place: resolution caps, clip length limits, aspect ratios, audio capabilities, and API token pricing — verified against ByteDance Seed and Volcengine Ark documentation as of May 2026.

Seedance 2.0 specs at a glance

Seedance 2.0 generates video at up to 1080p native resolution, with clips up to 12 seconds long, in five aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4). It supports text-to-video, image-to-video, and multi-reference conditioning. Audio is generated natively in a single pass: dual-channel stereo, with separate tracks for background music, sound effects, voiceover, and lip-synced dialogue. On the Volcengine Ark API, pricing is approximately 46 CNY per million tokens for pure generation (~$0.14/second effective rate).

Output specifications

Verified against ByteDance Seed launch post (Feb 12, 2026) and Volcengine Ark API documentation.

SpecValueNote
Max output resolution1080p nativeUp from 720p in 1.x
Max clip duration12 secondsUp from 8 s in 1.0; Lite was 5 s
Min clip duration4 seconds
Aspect ratios16:9, 9:16, 1:1, 4:3, 3:4
Frame rate24 fps (standard cinematic)Per Volcengine Ark docs
ModalitiesText-to-video (T2V), Image-to-video (I2V)Multi-reference I2V also supported
Multi-reference conditioningUp to 3 reference imagesCharacter + setting + style frame simultaneously
Multi-shot outputSupported — scene cuts within a single generation
WatermarkNone on paid API / hosted plans; Doubao free tier may apply branding

Specs reflect the Seedance 2.0 launch configuration (Feb 12, 2026). ByteDance may update limits between model sub-versions. Always verify against current Volcengine Ark documentation before building production integrations.

Native audio specifications

Audio generation is built into Seedance 2.0 as a standard feature — not a Pro add-on. All audio tracks are generated in a single model pass alongside the video.

Audio specValue
Audio formatDual-channel stereo
Background musicYes — prompt-driven, synchronized to visual rhythm
Sound effects (SFX)Yes — ambient and action-triggered
VoiceoverYes — prompt-driven narration track
Lip-syncYes — precise dialogue synchronization for characters
Multi-track outputYes — parallel tracks for music, SFX, voice
Audio in 1.x lineAdded in 1.5 Pro (late 2025); standardized in 2.0

API pricing (Volcengine Ark)

Pricing is per-token on the Volcengine Ark platform. The numbers below are sourced from TechNode's March 5, 2026 pricing report, corroborated by AICost and PANews. Token counts per clip are approximate and vary with resolution and duration.

Pricing specValueNote
Pure generation rate~46 CNY / million tokens≈ $6.35/M tokens at 7.24 CNY/USD
Effective cost per second~$0.14 / secondBased on ~308,880 tokens per 15-second clip
Video-editing mode rate~28 CNY / million tokens~39% cheaper than pure generation
New-user free quota5 million tokens freeFor new Volcengine Ark accounts; terms may change
Billing modelPay-per-token; no minimum commitment
API public beta openedApril 2, 2026

Source: TechNode, "ByteDance's Seedance 2.0 video model costs about $0.14 per second," March 5, 2026. Corroborated by AICost and PANews. Verified May 26, 2026.

Specs by access channel

The same underlying Seedance 2.0 model is accessible through three channels. The capability surface is identical; what differs is pricing model, commercial-use rights, and access requirements.

SpecVolcengine Ark APIseedance2-video.com (this site)Doubao / Jimeng (CN)
Max resolution1080p1080p1080p
Max duration12 s12 s12 s
Native audioYesYesYes
Multi-referenceYesYesYes
Commercial licenseYes (paid usage)Yes (every paid plan)No (consumer terms)
Pricing~$0.14/second pay-per-tokenCredit packs $15–$129Free (CN phone required)
Access requirementVolcengine account (CN entity often required)Email + USD cardChinese phone number
WatermarkNoneNoneDoubao branding on free tier

Capability data verified May 26, 2026. Pricing and terms may change — verify against current Volcengine Ark documentation and operator terms.

Generate a clip at these specs

Test the specs yourself — 1080p, up to 12 seconds, native audio, in your browser.

People also ask

What resolution does Seedance 2.0 output?

Seedance 2.0 outputs at up to 1080p native resolution. This is an improvement over the 1.x line, which maxed at 720p (with Lite at 480p). All three main access channels — Volcengine Ark API, seedance2-video.com, and Doubao — deliver 1080p.

How long can a Seedance 2.0 clip be?

Seedance 2.0 clips can be up to 12 seconds long, with a minimum of 4 seconds. This compares to 8 seconds maximum in Seedance 1.0 (and 5 seconds in the Lite tier).

What aspect ratios does Seedance 2.0 support?

Seedance 2.0 supports five aspect ratios: 16:9 (landscape), 9:16 (vertical/mobile), 1:1 (square), 4:3, and 3:4. The 1.x line only supported 16:9 and 9:16.

Does Seedance 2.0 generate audio?

Yes. Seedance 2.0 generates audio natively in the same pass as the video — no separate audio step. Supported audio types: background music, sound effects, voiceover narration, and lip-synced dialogue. Output is dual-channel stereo with separate tracks.

How much does the Seedance 2.0 API cost?

On Volcengine Ark, pure video generation costs approximately 46 CNY per million tokens. A 15-second clip uses roughly 308,880 tokens, making the effective rate about $0.14 per second of output (per TechNode, March 5, 2026). Video-editing mode is about 39% cheaper.

Frequently asked questions about Seedance 2.0 specs

Are Seedance 2.0 specs the same across all access channels?

Yes. The Volcengine Ark API, seedance2-video.com hosted interface, and Doubao consumer app all run the same Seedance 2.0 model and deliver the same capability surface (1080p, 12 s, native audio, multi-reference). The differences are in pricing model, commercial-use rights, and sign-up requirements.

What is "multi-reference conditioning" in Seedance 2.0?

Multi-reference conditioning means you can provide up to three reference images simultaneously — for example, a character photo, a setting photo, and a style frame. The model uses all three as conditioning inputs during generation, enabling more consistent character-in-scene output.

Does Seedance 2.0 support 4K output?

No. Seedance 2.0 caps at 1080p native as of May 2026. 4K output is not documented in the Volcengine Ark specifications. Upscaling via a separate tool after generation is possible but not built in.

Can I generate videos longer than 12 seconds?

A single Seedance 2.0 generation is capped at 12 seconds. For longer content, the standard approach is to generate multiple clips and edit them together, which is how most production workflows use the model.

What is the frame rate of Seedance 2.0 output?

Seedance 2.0 outputs at 24 fps — standard cinematic frame rate — per Volcengine Ark documentation as of May 2026.

How do I access Seedance 2.0 specs in the API documentation?

The authoritative spec reference is the Volcengine Ark API documentation at volcengine.com/product/ark. ByteDance Seed's launch post (seed.bytedance.com) also documents the headline capabilities. Third-party sources (including this page) cross-reference those primary sources.

Sources

  • Seedance 2.0 was announced on February 12, 2026 with 1080p output, up to 12-second clips, native audio, and multi-reference conditioning as standard features.ByteDance Seed, 2026-02-12
  • The Volcengine Ark API documentation specifies aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4), frame rate (24 fps), and pay-per-token pricing for Seedance 2.0.Volcengine, 2026-04-02
  • Per TechNode (March 5, 2026), Seedance 2.0 API billing is ~46 CNY/M tokens for pure generation and ~28 CNY/M for video-editing mode; a 15-second clip costs ~308,880 tokens (~$0.14/second effective rate).TechNode, 2026-03-05

Related pages

This page is operated by Vividra Labs LLC (Delaware), an independent third-party integrator using the official Seedance 2.0 API via Volcengine Ark. We are not affiliated with ByteDance. Spec data is verified May 26, 2026 against primary sources — terms and capabilities may change.