BLACK FRIDAY

Black Friday: 50% off annual plans + 30% bonus credits on Pro

Get 50% OFF
Ends in --d --:--:--

Seedance 2.5 Flash API: Model ID, Price and Limits

Nov 17, 2026

Short answer: The Seedance 2.5 Flash API is on BytePlus ModelArk as dreamina-seedance-2-5-flash-261115. For a 5-second 16:9 clip with no video input, examples are about $0.07/s at 480p and $0.15/s at 720p. Flash does not support 1080p or 4K. A 5-second 720p job with video input is listed at $0.84–$3.25. Check this site’s live API catalog.

Updated: November 17, 2026

What is the Seedance 2.5 Flash model ID?

The ModelArk pricing page and the Seedance 2.5 series tutorial list Flash as dreamina-seedance-2-5-flash-261115. Seedance 2.5 itself is listed as dreamina-seedance-2-5-260628. Use the Flash ID only when you intend to call Flash; the two IDs are not interchangeable, and Flash does not inherit 1080p or 4K from Seedance 2.5.

BytePlus announced Flash on November 12, 2026 and wrote that the API would launch on ModelArk on Sunday, November 15. The ModelArk model-release list records the Flash ID as a new video-generation release in its November 2026 section. Model presence on ModelArk does not mean this website’s API already exposes every parameter. As of November 17, 2026, this site’s Seedance 2.5 API models and pricing page did not list the Flash model ID. Confirm the live catalog and the selected mode before you hard-code an ID. The broader API surface is documented in the Seedance 2.5 API docs.

What token rates does Flash use on ModelArk?

Flash token prices are $7.0 per million tokens without video input and $4.3 per million tokens with video input, for the 480p and 720p online-inference column. Those rates are USD per million tokens, not a flat per-second checkout price.

On the same table, the 1080p and 4K online-inference cells for Flash are “Not supported.” The far-right column is Offline inference, separate from the three online-resolution columns. Seedance 2.5 has a priced online 4K cell; its trailing “Not supported” also refers to Offline inference.

The lower with-video token rate does not make a with-video job cheaper overall. ModelArk states that a video generated with video input costs more than one generated without video input, because token consumption includes input duration plus output duration. For a USD-per-million-token rate, estimate USD cost as billed tokens × rate ÷ 1,000,000. Actual consumption follows usage.completion_tokens in the API response; that field is a token count, not a dollar amount. ModelArk also applies minimum token consumption to video-input requests in the Seedance 2.5 series. You are charged only for successfully generated videos; failed jobs such as content-moderation failures are not billed on that page’s rule.

How much is a 5-second Flash video without video input?

On ModelArk’s 16:9, 5-second, no-video-input examples, Flash is $0.34 per video at 480p (about $0.07 per second) and $0.76 per video at 720p (printed as about $0.15 per second). $0.76 ÷ 5 seconds equals $0.152 per second, which is the 720p text-to-video and image-to-video figure in the BytePlus post. 1080p and 4K are not supported for Flash in that example table.

Item (conditions below)Seedance 2.5 FlashSeedance 2.5
Model IDdreamina-seedance-2-5-flash-261115dreamina-seedance-2-5-260628
Token rate, 480p/720p online, no video input$7.0 per million tokens$10.70 per million tokens
Token rate, 480p/720p online, with video input$4.3 per million tokens$6.40 per million tokens
5-second 16:9, no video input, 480p / 720p$0.34 / $0.76 (about $0.07/s and $0.15/s)$0.514 / $1.156 (about $0.103/s and $0.231/s)
5-second 16:9, with video input (input 2–30s), 480p / 720p$0.37–$1.45 / $0.84–$3.25$0.553–$2.152 / $1.244–$4.838
5-second 16:9, no video input, 1080p / 4KNot supported$2.843 / $4.568

Last verified: November 17, 2026, against the BytePlus ModelArk pricing page. These are ModelArk estimates under the stated conditions, not this site’s credit quote.

For comparison, ($0.231 − $0.152) ÷ $0.231 is about 34%, matching the “approximately 34% lower” wording for 720p in the BytePlus post. For this site’s request-level quoting behavior, see Seedance 2.5 API pricing and limits and the BytePlus Seedance 2.5 API setup notes.

What is the price range when the request includes video input?

For a 5-second 16:9 Flash output at 720p with video input, ModelArk lists $0.84–$3.25. The low end is a 2–4 second input; the high end is a 30-second input. The 480p with-video band is $0.37–$1.45 under the same output length and aspect ratio. 1080p and 4K remain not supported.

That band is higher than the no-video-input 720p example of $0.76, even though the with-video token rate ($4.3 per million) is lower than the no-video rate ($7.0 per million). Budget from the example range and from the live response, not from the token rate alone. If you are sending a still rather than a video file, use the no-video-input examples unless the request actually includes video. Related input-path writing is in Seedance 2.5 text to video.

Which resolutions, durations, and reference limits apply?

Flash supports 480p and 720p only. The tutorial’s capability overview lists 4–30 seconds of output and up to 20 reference assets (10 images, 5 videos, and 5 audio files). Individual input modes and reference types have their own requirements; the maximum file counts are not a guarantee for every request combination. Flash does not support 1080p, 4K, or Draft mode.

The series tutorial gives Seedance 2.5 a 50-asset cap (30 images, 10 videos, and 10 audio files) and lists 1080p and 4K (10-bit) in addition to 480p and 720p, with the same 4–30 second window. The ModelArk overview banner states Flash is available for videos up to 30 seconds at 480p/720p, with up to 10 images, 5 videos, and 5 audio clips as references. If you need 1080p, 4K, Draft mode, or more than 20 references, call Seedance 2.5 rather than Flash. Production-readiness checks for this site’s own API are in the Seedance 2.5 website API guide.

Does this site’s API already expose every Flash parameter?

This site’s public models-pricing page did not list Flash on November 17, 2026. ModelArk’s listing confirms its own model ID and billed sizes, not availability or credit prices on this website. Check the live catalog for the platform you will actually call; Flash does not acquire a 4K output option by being listed elsewhere.

When you call Flash on a platform that lists it, use that platform’s documented model ID and 480p or 720p settings. Use the applicable token price together with billed consumption to estimate the charge. BytePlus describes Flash as nearly twice as fast as Seedance 2.5, with results close to the flagship model; this is not a latency measurement from this site. This website’s Seedance 2.5 service and ModelArk maintain separate catalogs and billing.

What are the common questions about the Seedance 2.5 Flash API?

What model ID should you send?

dreamina-seedance-2-5-flash-261115 for Flash, and dreamina-seedance-2-5-260628 for Seedance 2.5. Confirm the ID against the live catalog you are calling.

Does a cheaper with-video token rate mean a cheaper video?

No. The with-video token rate is lower, but ModelArk’s examples and billing note show a higher total when video input is present, because input seconds are counted.

Can the Flash API return 4K?

No. Flash 1080p and 4K cells are “Not supported” on the ModelArk pricing page. Seedance 2.5 is the model that lists 4K there.

Are you billed if generation fails?

The ModelArk billing rules say you are charged only for successfully generated videos, and that failures such as content moderation are not billed. Follow the live account policy for the platform you use.

Try Seedance 2.5.

Seedance Team

Seedance Team