AI Models & Generation

Seedance 2.0 Mini vs Fast vs Full vs 2.5: Limits and ChatCut Controls

ChatCut Team
6 min read
Contents
  1. Last verified: October 28, 2026
  2. Which reference modes differ?
  3. Is Mini or Fast the same as Draft?
  4. How do matched prices compare?
  5. Choose a version for a test you can review
  6. Verification scope

Seedance 2.0 Mini and Fast both cover 4–15 seconds at 480p or 720p in ChatCut’s inspected integration. Full 2.0 adds 1080p here; its documented native 4K is blocked in ChatCut pending pricing. Seedance 2.5 extends the documented range to 30 seconds and adds native Draft controls and an explicit API selector for edit/extend tasks, with important integration gaps. These specifications do not establish a quality or speed winner.

Last verified: October 28, 2026

Official documentation was accessed October 27; ChatCut source and saved route checks were reviewed October 28. The ChatCut column describes the inspected product build, not a fresh production-generation test. See verification scope.

Version and public Ark model IDDocumented outputChatCut integrationSource
Full 2.0 — doubao-seedance-2-0-2601284–15s; 480p, 720p, 1080p, 4K480p/720p/1080p; 4K blocked because pricing is not wiredOfficial API, Ark tutorial, ChatCut audit
Fast — doubao-seedance-2-0-fast-2601284–15s; 480p/720pThose two resolution options; successful boundary jobs not established by this auditOfficial API, ChatCut audit
Mini — doubao-seedance-2-0-mini-2606154–15s; 480p/720pBoth options, default 720p; Mini is not 720p-onlyOfficial API, ChatCut audit
2.5 — doubao-seedance-2-5-2606284–30s; 480p/720p/1080pExposes generation and edit/extend requests; no Draft input; endpoint completion unverified in these probesOfficial API, ChatCut audit

BytePlus uses the corresponding dreamina- model IDs. A public model ID identifies a version; it does not establish entitlement on ChatCut’s configured Ark route.

Which reference modes differ?

Input or controlFull 2.0 / Fast / Mini2.5Source
FramesFirst frame, or first plus last; no mixing frames with reference modeSame separationOfficial API
Reference countsUp to 9 images, 3 videos, 3 audios; audio needs an image or videoUp to 30 images, 10 videos, 10 audios, 50 combined; audio-only references documentedOfficial API
Reference durationVideo/audio: 2–15s each, up to 15s combined per typeVideo/audio: 2–30s each, up to 30s combined per typeOfficial API
Shot timingOrdered shot descriptions; timestamps are not guaranteed controlsNative timestamp following documented; exact timing still needs reviewOfficial API, ChatCut audit
SoundNative audio switch exists; this ChatCut tool has no toggleSame tool gapOfficial API, ChatCut audit

Counts alone do not make a valid request: aggregate media duration, file constraints, portrait authorization and moderation still apply. Read generation references before combining media.

Is Mini or Fast the same as Draft?

No. Mini and Fast are model variants. Seedance 2.5’s native Draft workflow uses a 480p preview and a draft task to finalize at 1080p. ChatCut’s inspected submission tool does not forward the native Draft inputs, so Draft is not exposed on ChatCut yet. A lower-resolution normal generation is not that API workflow. Native finalization is described as similar rather than frame-identical. Source: official 2.5 tutorial, accessed October 27, 2026.

Official documentation also lists edit and extend workflows for Full 2.0, Fast and Mini. The explicit native omni_reference_task_type selector is documented for 2.5. Sources: official capability table, task API.

ChatCut exposes 2.5 edit and extend request modes, but the inspected adapter relies on prompt classification rather than forwarding the explicit native task selector. The Ark checks timed out. Treat those controls as exposed requests with unverified endpoint completion, not a demonstrated edit or extension result.

How do matched prices compare?

Use the confirmation quote for the exact model, resolution, aspect ratio, output duration and reference-video duration. Token billing can include input-video duration and minimums; settled credits can differ from an estimate. Current policy belongs to Plans and Credits, with estimate-versus-ledger guidance in generation credits.

No current matched four-model quote or quality/speed benchmark was verified for this article. The inspected Fast/Mini launch promotion has expired; it is not a current discount. A published BytePlus tariff is not automatically the tariff on ChatCut’s Ark route. No Seedance 4K price is available in the inspected ChatCut contract. Sources: ChatCut audit, provider pricing.

Choose a version for a test you can review

Your requirementCandidate to inspectWhat still needs checking
A 480p/720p clip within 15 secondsFull 2.0, Fast or MiniCurrent quote, route availability, measured waiting time and visual result
1080p within 15 secondsFull 2.0 or 2.5Exact output dimensions and whether the subject/motion passes your brief
More than 15 seconds, up to 302.5Reference duration budget, continuity and exact cuts
Native 4K in ChatCutNo verified option in this comparisonFull 2.0 model supports it; ChatCut pricing integration is pending
A native Draft-to-final workflow2.5 API outside this inspected toolChatCut has no Draft control yet
  1. Start with your own authorized reference assets and one fixed brief.
  2. Pick a common supported resolution, ratio and duration when comparing versions. Keep reference inputs identical.
  3. Review the model and credit estimate before approving each generation.
  4. Record elapsed time and actual ledger entries. Inspect the full take for motion, identity, sound and cuts.
  5. Keep the usable take on the timeline, then export and inspect the file. Generation dimensions and export dimensions are separate evidence.

For a different family, read Seedance vs Kling. The model capability matrix and failure guide explain route checks.

Frequently asked questions

What changes in resolution and duration?
Mini and Fast cover 4–15s at 480p/720p. Full 2.0 adds 1080p in ChatCut and documents native 4K, which ChatCut blocks pending pricing. 2.5 covers 4–30s up to 1080p. These are documented/exposed options, not successful boundary tests from this audit.
Which controls are exposed in ChatCut?
The inspected tool exposes version, duration, resolution and permitted frame/reference inputs, plus 2.5 edit/extend requests. Native Draft and an audio-off switch are not exposed. Endpoint success for the audited Ark probes remains unverified.
How do matched prices compare?
No current matched quote was verified. Compare confirmation quotes at identical supported settings, include reference-video duration, and use settled Credits History entries for actual cost. Expired launch promotions are not current prices.

Verification scope

Last verified 2026-10-28. Native specifications: BytePlus task API, 2.5 tutorial and Ark 2.0 tutorial, accessed 2026-10-27. ChatCut integration: internal source/receipt audit of product beta b3cbb24d, reviewed 2026-10-28; 4K pricing rationale. The configured Ark endpoint was unreachable from the test machine. Its 4K acceptance and the tested boundary requests remain unverified; no provider rejection or successful live generation is inferred from that timeout. No new paid test was run for this article.

Checking your footage...

Less editing. More creating.

Sign up free and make your first video today.

Try it for free