What is confirmed · What is not · What you can run right now
ByteDance's next video model has been shown publicly and is appearing inside their own apps. No API provider we buy from sells it yet, so this page does two things: says exactly what is known, and gives you a generator that works today.
These are Seedance-generated samples from our library, shown so you can see the family's output rather than a marketing reel. They are not 2.5 clips and we will not label them as such — when we can generate on it, the samples on this page get replaced with runs we did ourselves.
Drag to compare before and after
It is ByteDance's newest video model, publicly demonstrated and surfacing in their consumer apps, but not yet offered for sale through the API providers this site buys capacity from.
The distinction that matters here is between a model existing and a model being purchasable. Those are separated by weeks or months, and most pages published in that gap quietly blur them — describing capabilities in the present tense next to a generate button wired to something else entirely. What has been demonstrated publicly centres on longer continuous shots and multi-asset reference handling.
We have not run it, so we are not going to characterise how well any of that works.
What we did do is check the shelf rather than guess. On 1 August 2026 we pulled fresh catalogues from both providers we buy through: one carries Seedance 2.0, 2.0 Fast, 2.0 Mini and 1.5 Pro but no 2.5 entry at all; the other lists a 2.5 row marked as coming soon with no price against it.
A listing without a rate is not something you can call. That is the whole basis for everything on this page, and it is the kind of claim we would rather be checkably wrong about than vague.
So the honest summary: if you came here to use Seedance 2.5 today, you cannot, here or anywhere that sells API capacity. What you can do is run the generator below on the current release, which is genuinely good and available on free credits — or read the tested write-up on the Seedance 2.0 model page before you spend any of them.
Creative engine
Seedance 2.0 through our provider, from a text prompt or a still image.
Provider catalogues pulled fresh the day this page shipped. When that changes, this page changes.
Numbers we have not verified against a primary source or our own runs are not printed here as facts.
The generator below is live. This is how to get something usable out of it.
Camera move, subject, light, pace — "slow push-in on a surfer paddling out through glassy dawn water, low sun flaring across the lens". Video models reward shot direction and punish plot summaries.
Feeding a still fixes the look before any motion is invented, which removes most of the variance between runs. It is the single biggest quality lever available on any current video model.
Change the camera move or the pace, not both. Video runs cost more than image runs, so a disciplined comparison gets you to a keeper in fewer generations.
Two of them are answerable today.
You saw the demos circulating and want to know if you can actually use it. Answer: not through any API provider yet, and this page updates when that stops being true.
The version number matters less than getting the shot. The studio below runs the current release on free credits — start there and stop waiting.
Deciding where to spend generations. Read the tested notes on Seedance 2.0 and Veo 3.1, or take a still and extend a clip you already have instead of starting over.
Answered from what we can actually check.
No. The generator on this page runs Seedance 2.0, and it says so in the interface as well as here. We wire a model when a provider sells it; as of 1 August 2026 neither of ours does, so there is nothing to wire. Anyone offering you 2.5 API access today is worth a second look.
We do not know, and neither does anyone quoting a date without a source. One of our providers lists it as coming soon, which is a statement of intent rather than a schedule. This page changes the day it becomes callable — that is the only commitment worth making.
Public demonstrations have focused on longer continuous shots and handling multiple reference assets in one generation. We are deliberately not repeating specific numbers for duration, resolution or frame rate: we have not run the model, and pre-release figures get revised. When we can test it, this section gets measurements instead of description.
Yes, and it is what we would use ourselves today. It is the current shipping release, it is available at several quality tiers so you can trade cost against resolution, and there are tested notes on its model page rather than launch-day impressions. New accounts get free credits, which is enough to find out whether it suits your shot.
Prompts written as shot direction — camera, subject, light, pace — tend to transfer across video models, because they describe what you want rather than exploiting one model's quirks. Prompts tuned to a specific model's failure modes usually do not. Writing the former is the better habit either way.
Because the honest version of this page is more useful than the alternatives on offer, which are either silence or an implied generator that does not exist. You get a checked status, a working studio on the current release, and a page that will still be here — and correct — on the day access opens.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
One-time top-ups — buy extra credits any time you run low.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
Charges appear as “KAVEL AI” on your card statement. You can cancel any time from Settings → Billing; cancellation takes effect at the end of the period you already paid for.
Powered by
Use one balance across the supported image and video models (Nano Banana 2 / Nano Banana Pro / Seedance 2 / Veo 3.1), and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. Mini is cheaper still. |
| Video | Veo 3.1 | Billed per video (Lite). About 34 credits at 1080p, about 23 at 720p. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. About 8 credits per image. Start here to test. |
The studio below runs today's release on free credits — and when 2.5 opens up, this page will be the first to say so.