Seedance 2.0: The Flagship Text & Image-to-Video Model

Seedance 2.0 is the flagship model in the Seedance Model Suite — a single unified engine that handles both Text-to-Video and Image-to-Video generation.

It ships inside Imagine Dance Studio with a 40-credit default job, which is the sweet spot for: product ads, character moments, lifestyle spots, and client pitch visuals that need natural human motion and film-grade lighting.

Try the 40-credit flagship generation in the Hero Studio on the Imagine Dance landing page.

When to choose Seedance 2.0

Pick Seedance 2.0 whenever the brief asks for quality first. It runs slower than Seedance 2.0 Fast, but consistently outperforms on:

• Physical realism — correct limb anatomy, no warping, plausible skeletal motion.
• Reference image identity lock — faces, outfits, and product shapes hold across the whole clip.
• Prompt adherence — cinematography terms (dolly-in, rack focus, magic hour) translate directly into shot language.
• Native audio pairing — Seedance 2.0 is the model you want for "sound-on" first-drafters.

See real output samples of Seedance 2.0 in the Demo Reel section on the landing page.

Text-to-Video: prompt structure that works

The model responds best to a three-part formula. Paste this template every time:

Subject + action + environment. Cinematography: [shot / camera move / lens]. Lighting: [key light + tone]. Mood: [feeling], 8s, 16:9, 720p.

Experiment directly with Text-to-Video mode from the landing page Studio at Imagine Dance — Text-to-Video is the first segmented toggle on the left.

Example — a lifestyle product ad

A young designer opens a matte-black notebook at a wooden table with warm coffee steam. Cinematography: close-up rack focus from fingers to embossed cover, gentle parallax dolly. Lighting: soft north window light, warm 3200K practical on the side. Mood: calm focused ritual, 5s, 16:9, 720p.

Keep subject, action, and environment in the first sentence. Cinematography and lighting as second and third. Seedance 2.0 parses punctuation lightly — commas are better than newlines in the prompt box.

Image-to-Video: reference first

Drop a reference image before typing anything. Seedance 2.0 treats the image as a frame-zero anchor, so:

1. Characters keep their face, hair, and wardrobe — even when you add motion prompts.
2. Products retain logos, finish, and proportion across 5–8 second shots.
3. Background color grading bleeds into the rest of the clip naturally.

Switch to Image-to-Video mode and upload your frame-zero reference inside the Hero Studio on Imagine Dance before writing the prompt.

Best prompts to pair with a reference

• Keep the environment sentence short. The reference already defines it.
• Push motion, camera, and pacing instead: "slow cinematic zoom out, subtle hair and fabric breeze."
• Avoid re-describing the person or product. Let the image do that work.

Compare all 4 family models side-by-side in the Powered by Seedance Family section of the landing page.

Recommended settings per creative job

| Job type | Ratio | Duration | Resolution | Credits | |---|---|---|---|---| | Social feed ad | 16:9 | 5s | 720p | 40 | | TikTok / Reels | 9:16 | 8s | 480p | 40 | | Client pitch frame | 1:1 | 4s | 720p | 40 | | Youtube 16:9 hero | 16:9 | 8s | 720p | 40 |

Workflow tip: draft in Seedance 2.0 Fast first

For teams shipping volume, a cheap 20-credit Seedance 2.0 Fast exploratory draft → then Seedance 2.0 final pass beats iterating on the flagship every time.

Learn how to draft → finalize in our full guide: Seedance 2.0 Fast: Under 5s Drafts & Storyboard Loops.

For multi-character and artistic-style scenes, read Seedance 2.5: Director-Level Multi-Character & Style Control.

Ready to produce your first 40-credit flagship shot? Open Imagine Dance and generate.