Seedance 2.5

Seedance 2.5 brings continuous 30-second shots, more references, and easier edits. Available in ChatCut web, ChatGPT, Claude Code, and the WorkBuddy plugin. Choose a plan below to get started.

Upgrade Now

What does Seedance 2.5 support?

As reviewed on 28 October 2026, Seedance 2.5 documents 4–30-second output at 480p, 720p or 1080p, plus a 30-image / 10-video / 10-audio reference budget. The inspected ChatCut tool exposes generation, editing and extension controls, but no Draft input. Current Ark boundary and mode probes timed out; those checks remain unverified, distinct from a separate local 30-second previs result.

Model / public IDCapabilities, integration and evidence
Seedance 2.5
doubao-seedance-2-5-260628

Availability: available_in_inspected_build
Restriction reason: unknown
Endpoint verification: transport_inconclusive

Pinned integration and saved dev checks; production unverified. Source links describe native specifications.

Reviewed:

Native documentation: 4–30s; 480p/720p/1080p; up to 30 image / 10 video / 10 audio references; native edit, extend and Draft.
Native documentation; not an endpoint test; . Native specification source 1 (2026-10-27) Native specification source 2 (2026-10-27)

ChatCut integration: Generate/Edit/Extend controls; 30/10/10 local caps. No Draft input or output-audio toggle in inspected tool.
Pinned integration; production unverified; . Native specification source 1 (2026-10-27) Native specification source 2 (2026-10-27)

Endpoint evidence: Ark boundary and explicit edit/extend semantics unverified after timeouts. Separate local previs run exported 30s.
Saved development evidence; not production certification; . Native specification source 1 (2026-10-27) Native specification source 2 (2026-10-27)

How are 50 references counted?

The Ark public ID is doubao-seedance-2-5-260628; BytePlus uses dreamina-seedance-2-5-260628. The 50-file budget is 30 images + 10 videos + 10 audios, not 50 of any type. Native audio-only references are documented. Video and audio references are each 2–30 seconds, with 30 seconds aggregate per type. Frame inputs cannot mix with reference mode; a last frame requires a first frame.

Which edit, extend and Draft controls are exposed?

The native API documents explicit reference/edit/extend selectors and Draft-to-final. In the inspected ChatCut integration, Edit requires one source video and a prompt; Extend takes one source and an added duration. The adapter uses prompt classification rather than forwarding the native explicit task selector. Its native selector behavior and extension semantics remain unverified after Ark timeouts. There is no draft or draft-task input in the inspected tool.

Native edit sources are 4–30 seconds. Edit requires adaptive ratio and duration; Extend requires adaptive ratio. Native Draft uses a 480p preview followed by a 1080p final from that draft task; it is distinct from Seedance Mini. Integer-second timestamp prompting belongs to 2.5, without guaranteeing frame-exact cuts. The checked output enum stops at 1080p; full 2.0 has separate native 4K documentation.

Make a short reference-led sequence

  1. Choose Seedance 2.5 in the current Video Gen model list and begin with a five-second 16:9, 480p text or first-frame request.
  2. For reference mode, label each image, video or audio role and check both counts and total media durations. For Edit or Extend, use that mode's single source slot.
  3. Review the current estimate, submit, then inspect completed cuts, subject identity and audio before trimming and assembling on the timeline.

A local previs acceptance run on 27 October exported 30 seconds and passed 15 of 17 checks. Side-view proof failed and speaker/no-music audio remained unresolved. This is a bounded workflow result, not proof of all native controls, route boundaries or production availability.

Credits and next steps

Current rates belong to Credits Policy. Check the preflight quote with the exact provider, model, cohort, resolution, ratio, duration and reference-video duration; token settlement and Agent usage can affect total spend. This review does not establish a current quote or a quality winner.

See the model matrix, full 2.0 limits, Fast and Mini, reference guide and failure recovery.

Last fact review: 28 October 2026. Official sources were accessed on 27 October; integration contracts and saved development checks were reviewed on 28 October. This is dated evidence, not a live certification of every production account. No new paid generation was performed for this page.

Turn Everything to Video

Product Ads selected.

More control betweenthe idea and the cuts

Seedance brings the motion. ChatCut keeps the references, revisions, and final edit in one continuous workflow.

Change the shotwithout leavingthe cut

Ask for the adjustment where the video already lives. Review the result, keep the version you like, and continue editing on the same timeline instead of moving files between tools.

Direct the destination,not onlythe opening

Set where the shot begins and where it needs to land. The motion in between becomes a controlled transition rather than a generation that wanders away from the idea.

Turn separate beatsinto one continuousmove

Use a sequence of visual moments to establish the route through a scene. Camera motion, subject movement, and pacing connect into a shot that reads as one idea instead of separate clips.

Keep the castreadable as thescene unfolds

Give every character a clear role and let the action carry the story. Appearance, staging, and personality stay legible long enough for the scene to feel performed, not assembled.

Start with the visualyou alreadytrust

Bring a reference into the conversation when words are not enough. ChatCut keeps the source beside the generation, so the direction stays visible from the first prompt to the final frame.

Generate AI Videos With Full Editing Control

Generate video from text, images, or a first and last frame, then edit, caption, voice, and score it, all in one place.

Generate a video

Image to video

First & last frame

Camera moves

AI voiceover

AI music

Captions

Motion graphics

Generate cinematic videos from a single prompt

Describe the shot, motion, and timing — then generate up to 30 seconds and continue the cut in the same editor.

Seedance 2.5 00:30

4–30-second generation

Request up to 30 seconds, then review cuts, character identity, camera motion, and sound.

Describe the video you want to create…

Seedance 2.5 30s 16:9

Prompt to Video

Describe a complete shot and generate a ready-to-edit video directly into your timeline.

Reference Direction

Guide motion, composition, and character identity with images, clips, or style frames.

Motion Guidance

Control camera movement, subject action, and transitions in plain language.

Timeline Editing

Trim, extend, revise, and combine every generated shot without leaving the editor.

Subscribe to get your AI video editor

Unlock member features, models, and export allowances. Get more credits every month.

$0

No credit card required

Free starter allowance

Try for free
  • Basic editing and exports
  • AI images and motion graphics
  • AI captions and music
  • AI avatars; voice cloning and previews
  • Web export limit: 60 minutes total
CodexClaudeWorkBuddy

Some features are completely free in Codex, Claude, and WorkBuddy.

Learn More
$21 $25 /month 16% off

$252 · Billed annually

100 credits / month

Get started
  • Seedance 2.5 (up to 208s of video)
  • Seedance 2.0 (up to 294s of video)
  • MiniMax H3 (Up to 256s of video)
  • MiniMax H3 Max (Up to 416s of video)
  • GPT Image 2 (up to 384 images)
  • Nano Banana 2 (up to 256 images)
  • Motion graphics (up to 500 clips)
  • AI music (up to 454 songs)
  • Clone voices and reuse them in projects (up to 1 voices)
  • All AI agent editing capabilities
  • Motion graphics transparent export
  • No export limit
$38 $50 /month 24% off

$456 · Billed annually

200 credits / month

Get started
  • Seedance 2.5 (up to 416s of video)
  • Seedance 2.0 (up to 588s of video)
  • MiniMax H3 (Up to 512s of video)
  • MiniMax H3 Max (Up to 833s of video)
  • GPT Image 2 (up to 769 images)
  • Nano Banana 2 (up to 512 images)
  • Motion graphics (up to 1,000 clips)
  • AI music (up to 909 songs)
  • Clone voices and reuse them in projects (up to 2 voices)
  • All AI agent editing capabilities
  • Motion graphics transparent export
  • No export limit
$141 $200 /month 30% off

$1692 · Billed annually

800 credits / month

Get started
  • 400 credits / month
  • 800 credits / month
  • 1,200 credits / month
  • 2,000 credits / month
  • Seedance 2.5 (up to 1,666s of video)
  • Seedance 2.0 (up to 2,352s of video)
  • MiniMax H3 (Up to 2,051s of video)
  • MiniMax H3 Max (Up to 3,333s of video)
  • GPT Image 2 (up to 3,076 images)
  • Nano Banana 2 (up to 2,051 images)
  • Motion graphics (up to 4,000 clips)
  • AI music (up to 3,636 songs)
  • Clone voices and reuse them in projects (up to 8 voices)
  • All AI agent editing capabilities
  • Motion graphics transparent export
  • No export limit

Seedance 2.5 questions

Need more help? team@chatcut.io

Checking your footage...

Less editing. More creating.

Sign up free and make your first video today.