Sume vs Seedance: BytePlus ModelArk direct, or one API across models
Seedance 2.5 is ByteDance's video model, sold by the token on BytePlus ModelArk. Sume offers Seedance in one multi-model API and agent. Access, limits, pricing.

Seedance is ByteDance's video generation model family. Developers call it on BytePlus ModelArk and pay per token; consumers can use it in Dreamina. Sume is a video agent platform that offers Seedance as part of its video catalog (seedance-2.5, seedance-2-mini, seedance-2, seedance-2-fast), next to other model families. So the practical question is whether to call ModelArk directly or go through Sume.
Seedance facts come from ByteDance and BytePlus pages, read on 2026-09-24. Sume facts come from Sume's docs, pricing page, and public catalog.
What is Seedance?
ByteDance describes the current flagship this way: “Seedance 2.5 is a next-generation audio-video joint generation model, built for 30-second storytelling with precise reference control and powerful editing capabilities.” Its models page links both a playground and Dreamina.
- Access: “You can call the API and try the model online on the BytePlus ModelArk platform” (API guide). The API uses async tasks, API-key authentication, and an optional callback URL for status changes.
- Versions on the API: Seedance 2.5; 2.0, 2.0 fast, and 2.0 mini; 1.5 pro; and 1.0 pro (pricing).
- Limits: Seedance 2.5 makes 4–30 second clips at 480p, 720p, or 1080p; Seedance 2.0 makes 4–15 second clips and goes up to 4K; generated audio is mono (API guide).
- Inputs: Seedance 2.5 supports 11 languages and “up to 50 omni reference assets per request” (quick start).
- Getting started: the quick start recommends a BytePlus balance above USD 30 to activate the model, and video URLs are “Retained for 24 hours” (quick start).
Can I use Seedance through Sume?
Yes. Send one of seedance-2.5, seedance-2-mini, seedance-2, seedance-2-fast as model to Sume's OpenRouter-compatible POST /v1/videos, or send sume/auto and let Sume choose the family. Sume's video docs say seedance-2.5 accepts 4–30 seconds at 480p, 720p, or 1080p, and that seedance-2 also offers 1080p.
What is Sume?
Sume is a video agent platform: people brief an agent in the Agents tab, save the result as a reusable Format, and call that Format, the agent, or individual models from their own backend over one API, a hosted MCP server, a CLI, and a TypeScript SDK, all billed from one wallet.
- Agents: a chat workspace at sume.com/agents where the agent plans a video, asks before it spends, and generates stills, avatars, voice, lip-sync, and a finished timeline.
- Formats: a saved recipe your backend calls with
POST /v1/formats/{handle}/{slug}/runs, plus 27 ready-made Formats at thesumehandle. See What is a Sume Format?. - Models: an OpenRouter-compatible
POST /v1/videosendpoint over a catalog of video models, withsume/autorouting, plus image, music, speech, and Avatar 1.0. - Avatar 1.0: create a reusable avatar ($0.95 per avatar) and generate 4–60 second talking videos at $0.184/s standard, $0.245/s plus, $0.55/s max (no product image).
- Developer surfaces: the Developer API, a hosted MCP server for Claude Code, Cursor, and Codex, a CLI, and a TypeScript SDK.
- Plans are Free $0, Pro $40 / month, Startup $120 / month, Scale $400 / month. Usage draws from one prepaid wallet at each model's published USD rate plus a 5.5% agent fee by default; routed video models are priced at the provider's list price × 1.25. The public catalog states both.
How do Sume and Seedance on ModelArk compare?
| Compared | Sume | Seedance (BytePlus ModelArk) |
|---|---|---|
| Models | seedance-2.5, seedance-2-mini, seedance-2, seedance-2-fast, kling-3, wan-3.0, grok-imagine-video-1.5, minimax-h3, minimax-h3-max, gemini-omni-flash-1.1, h3-max-recast, or sume/auto to let Sume pick | Seedance 2.5, 2.0 (fast, mini), 1.5 pro, 1.0 pro |
| Seedance 2.5 limits | seedance-2.5: 4–30 seconds at 480p, 720p, or 1080p | 4–30 seconds at 480p, 720p, or 1080p |
| API | REST API at api.sume.com with a Bearer key; async jobs you poll or take by signed webhook; Idempotency-Key on creates | Async tasks at ark.ap-southeast.bytepluses.com; API-key auth; optional status callback |
| Billing | One prepaid wallet; each model's published USD rate plus a 5.5% agent fee; routed video models at the provider's list price × 1.25 | Per million tokens; only successfully generated videos are billed; prepaid token packs available |
| Output storage | Sume-hosted files on media.sume.com | Video URLs retained for 24 hours |
What does Seedance cost?
ModelArk prices Seedance by the token and publishes worked examples per clip. It also says: “You are only charged for successfully generated videos” (pricing).
- BytePlus also sells prepaid Seedance 2.5 token packs, for example “Light Plan | $32.00 / 5 million Tokens | valid for 3 months” (product page).
- On Sume,
seedance-2.5,seedance-2-mini,seedance-2,seedance-2-fastare billed at the provider's list price × 1.25 plus the 5.5% agent fee, from the same wallet as every other model. Exact per-model rates are in the catalog and on API pricing.
| Seedance 2.5 on ModelArk | Price |
|---|---|
| 480p or 720p output, input without video | USD 10.70 per million tokens |
| 480p or 720p output, input with video | USD 6.40 per million tokens |
| 1080p output, input without video | USD 11.70 per million tokens |
| Example: 5 s, 720p, 16:9 | About USD 1.156 per video (0.231 per second) |
Should I call ModelArk directly or use Sume?
- Call ModelArk directly if Seedance is the only model family you need and you want BytePlus's token pricing, token packs, and its Seedance 1.x and 2.0 variants.
- Use Sume if Seedance is one of several models in a job: one API key, one wallet, and one job format across Seedance, Kling, Wan, MiniMax, Gemini, and Grok video, plus Avatar 1.0, voice, and timeline assembly, with outputs hosted on media.sume.com.
- Use Sume if you want an agent to plan the video and save the recipe as a Format your backend runs with an idempotency key, a spend cap, and an optional webhook.
How do I try Sume?
Start a thread in the Agents tab, or create an API key and follow the quick start. Plans and per-model rates are on Pricing and API pricing; How Sume pricing works explains the wallet.
Vendor details change often. Everything we say about other products here comes from their own pages as of 2026-09-24; check those pages before you decide.
Sources
- ByteDance Seed: Seedance 2.5
- ByteDance Seed: models
- BytePlus ModelArk: video generation API
- BytePlus ModelArk: Seedance pricing
- BytePlus ModelArk: Seedance quick start
- BytePlus Seedance product page
- Sume video generation (catalog limits)
- Sume basics
- Sume pricing
- Sume API pricing
- Sume public catalog (GET /v1/catalog)
- Sume Format API
- Sume hosted MCP
- Sume core concepts (jobs, media.sume.com, refunds)
- Sume generation admission (failed jobs are refunded)
Related posts
More in Comparisons
- Sume vs Synthesia: AI avatar video platforms compared
Synthesia makes long avatar videos for business communication; Sume is a video agent with a multi-model API. Plans, API access, MCP, and limits compared.
- ACE-Step 1.5 open-source music vs Lyria on Sume's Music Router
ACE-Step 1.5 is MIT-licensed, makes up to 10-minute songs and runs in under 4 GB of VRAM. How it compares with Lyria 3.5 via Sume's Music Router.
- Is Firefly output 'commercially safe'? Firefly vs partner models
Adobe says outputs from Firefly AI models are safe for commercial use and trained on licensed and public domain content. Partner models are listed separately.
- AI disclosure API fields: TikTok, Instagram, YouTube and Snap
The four field names that mark a video as AI-made: TikTok is_aigc, Instagram is_ai_generated, YouTube containsSyntheticMedia, Snap ai_content_source.
Written by Sume