Here is the straight answer: the Seedance 2.5 Flash API is priced but not open yet. ByteDance sells Seedance to developers through BytePlus ModelArk, and on November 12, 2026 ModelArk's price list added Flash, with its model ID, dreamina-seedance-2-5-flash-261115, at $7.00 per million tokens. The API itself still answers that the model doesn't exist. Build on Seedance 2.5 today, keep the model ID in one setting, and switching to Flash is a one-line change.
dreamina-seedance-2-5-flash-261115On ModelArk's price list; not accepting calls yetdreamina-seedance-2-5-260628On BytePlus ModelArk todayTwo ways to call Seedance from code
- BytePlus ModelArk, ByteDance's own API. You need a BytePlus account and API key, and you pay BytePlus per token.
- The Lumeta API, one key for Seedance and every other top image, video and audio model, paid in Lumeta credits. Lumeta is the cheapest place to run Seedance 2.5 (details at lumeta.ai/pricing#seedance), and it also works from Claude, ChatGPT and other AI assistants through its MCP server.
Seedance on BytePlus ModelArk
A Seedance video is a task: you create it, then check on it until the video is ready. This is the shape of a text-to-video request for Seedance 2.5. When Flash opens, change SEEDANCE_MODEL and nothing else (Flash stops at 720p, so keep resolution at 480p or 720p).
export ARK_API_KEY=your-byteplus-key
export SEEDANCE_MODEL=dreamina-seedance-2-5-260628 # Flash: dreamina-seedance-2-5-flash-261115, once ModelArk accepts it
curl https://ark.ap-southeast.bytepluses.com/api/v3/contents/generations/tasks \
-H "Authorization: Bearer $ARK_API_KEY" -H "Content-Type: application/json" \
-d '{
"model": "'"$SEEDANCE_MODEL"'",
"content": [{"type": "text", "text": "A courier in a yellow rain jacket cycles through a neon-lit street at night, slow push in."}],
"resolution": "720p",
"ratio": "16:9",
"duration": 5,
"generate_audio": true
}'
# returns {"id": "cgt-..."}; then GET .../contents/generations/tasks/cgt-... until status is "succeeded"
Reference images, videos and audio go in the same content list as image_url, video_url and audio_url items. Seedance 2.5 takes up to 30 images, 10 videos and 10 audio clips; Flash takes up to 10 images, 5 videos and 5 audio clips, in clips up to 30 seconds. Check BytePlus's ModelArk documentation for your region's endpoint and the current limits.
Seedance on the Lumeta API
One key, plain JSON, and a free cost estimate before you spend anything. Pick the engine in input:
export LUMETA_API_KEY=lm_live_...
# 1. What will it cost? (free)
curl https://api.lumeta.ai/v1/estimate \
-H "Authorization: Bearer $LUMETA_API_KEY" -H "Content-Type: application/json" \
-d '{"model": "seedance", "input": {"engine": "2.5", "prompt": "A courier cycles through a neon-lit street at night", "resolution": "720p", "duration": 5}}'
# 2. Make it
curl https://api.lumeta.ai/v1/generations \
-H "Authorization: Bearer $LUMETA_API_KEY" -H "Content-Type: application/json" \
-d '{"model": "seedance", "input": {"engine": "2.5", "prompt": "A courier cycles through a neon-lit street at night", "resolution": "720p", "duration": 5}}'
# 3. Wait for the video (use the id from step 2)
curl "https://api.lumeta.ai/v1/generations/gen_123?wait=45" -H "Authorization: Bearer $LUMETA_API_KEY"
Python, Node, webhooks and the full reference: Lumeta API docs. Flash will be added to the Lumeta API the day ByteDance releases it.
Get a Seedance API key in a minute
One key for Seedance 2.5, Seedance 2.0 and Flash on launch day, plus Veo, Kling and more.
How to spot Flash the moment it lands
- Keep the model ID in one setting, like
SEEDANCE_MODELabove, never hard-coded in several places. - Test the ID for free. Send a create request with
dreamina-seedance-2-5-flash-261115and a 1-secondduration, which every Seedance model refuses before making a video, so it costs nothing. An answer that the model does not exist means Flash isn't open to you yet; a complaint about the duration means it is. Watch ModelArk's model list in your BytePlus console too. - Or check this page. We post the day it opens here, with measured run times and the limits.
Seedance API pricing
ModelArk bills Seedance by tokens. A clip's tokens are roughly its width times height times frames, divided by 1,024, so a 5-second 720p clip is about 109,000 tokens. At BytePlus's list prices that is about $1.17 for Seedance 2.5, and $0.76 for both Seedance 2.5 Flash and Seedance 2.0, which share a price: $7.00 per million tokens, or $4.30 when the request includes a video. On Lumeta the same Seedance 2.5 clip is $0.73. See the pricing page for every resolution.
API questions
Is the Seedance 2.5 Flash API available?
Almost. Since November 12, 2026, BytePlus ModelArk's price list shows Flash with its model ID and price, but the API doesn't accept that ID yet: when we tried, it answered that the model does not exist. Our estimate is that it opens in mid to late November 2026. Seedance 2.5 works today, as dreamina-seedance-2-5-260628.
What is the Seedance 2.5 Flash model ID?
dreamina-seedance-2-5-flash-261115, as listed on BytePlus ModelArk's price list. Like other Seedance IDs, it ends in a version date, here November 15, 2026.
What is the cheapest Seedance API?
For Seedance 2.5, Lumeta: a 5-second 720p clip costs $0.73, against 59% more on Replicate, 123% more on WaveSpeed and about three times as much on fal (prices checked November 9, 2026). Details at lumeta.ai/pricing#seedance.
Can I use Seedance from Claude or ChatGPT?
Yes, through Lumeta's MCP server, which lets AI assistants make Seedance videos on your Lumeta account. Setup guides are on Lumeta's developer pages.
Sources: BytePlus ModelArk (model IDs and request format, as used in production on Lumeta), ModelArk's price list (Flash's ID and price), BytePlus's Seedance 2.5 announcement, Dreamina's Seedance 2.5 Flash page. Launch timing is our estimate.