Which Seedance model should a beginner use?
Start with Seedance 1.0 Lite when the goal is to test the prompt and framing at the lowest current Seedance charge.
Seedance guide
Facts checked August 12, 2026
How to use Seedance: pay for the answer you need, not the newest model
The Seedance family has different cost, resolution and workflow strengths. Start by naming the uncertainty in the shot, choose the cheapest model that can resolve it, and spend on the final tier only after the short take works.
Cheapest draft
18 credits
Seedance 1.0 Lite
2.0 first tier
121 credits
Fast · 720p · 5s
1080p option
62 credits
Seedance 1.0 Pro
Long option
30 seconds
2.5 · test short first
Choose the Seedance model by task
| Task | Start with | Why |
|---|---|---|
| Basic prompt and framing test | 1.0 Lite | Lowest Seedance charge in the catalog |
| Complex motion or interaction | 2.0 Fast | Verified 2.0 path before Standard |
| 1080p delivery | 1.0 Pro | Current Seedance 1080p option |
| Animate a source photo | 2.5 Image to Video | Current verified product path |
| Long selected scene | 2.5 · 30s | Available only after the five-second idea is locked |
The reusable six-step Seedance workflow
1. Choose the question
Decide whether you are testing a prompt, preserving an image, staging complex interaction, or producing a final high-resolution take.
2. Choose the cheapest model that can answer it
Use 1.0 Lite for a basic draft, 2.0 Fast for difficult motion, 1.0 Pro for 1080p, or 2.5 for its current text and photo modes.
3. Write observable motion
State subject, action, camera, timing, light, sound and constraints.
4. Run five seconds first
A short test exposes prompt and motion problems before the long or sharper take.
5. Change one variable
Keep the successful settings and edit only the part that missed.
6. Finish or download
Upgrade the selected shot, review the full output and preserve its prompt, model and settings.
Write motion, camera and timing—not a pile of adjectives
Subject and setting → action → camera → timing → light → sound → constraints.
- Use verbs and a visible end state.
- Choose one dominant camera move for a five-second draft.
- Quote exact dialogue and identify the speaker.
- Add no text, logos, cuts or extra people when those constraints matter.
Use the failure state to choose the next action
| Observed state | Next action |
|---|---|
| Blocked before submit | Remove the unsafe or deceptive request; no provider spend should occur. |
| Insufficient credits | Use Lite or purchase only the exact gap. |
| Provider failed or timed out | Verify the one-time credit return, then simplify. |
| Completed but motion missed | Change only action or camera, not the entire prompt. |
| Identity drift from an image | Use a clearer source and smaller subject motion. |
Go deeper by version or input mode
How to Use Seedance 2.0
Exact Fast/Standard workflow with paid canary timings.
Open 2.0 tutorialSeedance Image to Video
Source preparation, motion prompting and failure fixes.
Open image guideSeedance 2.0 Prompts
Six original recipes with per-recipe test status.
Open 2.0 promptsFrequently asked questions
How do I get better Seedance results?
Describe observable subject motion, one main camera move, timing, light, sound and constraints. Test five seconds and change one variable per iteration.
Should I always use the newest Seedance model?
No. Use 1.0 Lite for low-cost drafts, 1.0 Pro for 1080p, 2.0 for complex interaction, and 2.5 only when its specific modes answer the task.
How long does generation take?
Queue time varies. First Take Lab’s dated 2.0 canaries took about 125 seconds for Fast and 222 seconds for Standard; those are observations, not guarantees.
First Take Lab is independent and not affiliated with ByteDance. Model selection advice is based on the current First Take Lab catalog and dated first-party tests.