Seedream 4.0 is the entry tier of ByteDance's Seedream family. It is the one that takes a resolution rather than a quality switch, and the one cheap enough to run several times while you are still deciding what you want.
Four steps, and the last one is what this tier is actually for.
Choose the shape by name before writing — it changes what belongs in the frame.
What the thing is, and where the light comes from. Both do more than any adjective.
Resolution does not change the composition, and 4K is only worth generating once you like it.
At 20 credits, an attempt here costs less than a single image on the top tier.
Seedream 4.0 is the oldest tier of ByteDance's Seedream family still worth offering, sitting below 4.5, 5.0 Lite and 5.0 Pro. It is also the odd one out in how you talk to it: where the later tiers take an aspect ratio and a quality setting, 4.0 takes a named framing preset — square, portrait_16_9, landscape_21_9 and so on — plus a resolution of 1K, 2K or 4K.
Only the prompt is actually required. What it is for is the part of the process before you know what you want. Most of the images anyone generates are thrown away, and they are thrown away for reasons that have nothing to do with model tier: the composition was wrong, the idea did not work, the subject was in the wrong place.
Finding that out at twenty credits instead of fifty-five is the whole argument for this tier. Where it visibly gives up is text inside the image — 4.5 is the tier that fixed that — and faces under difficult light, which is what 5.0 Pro is for.
Creative engine
One flat rate, and the resolution you pick does not change it.
Below 4.5, 5.0 Lite and 5.0 Pro, and the cheapest way to run the same prompt.
square, portrait_4_3, landscape_16_9 and six more, plus a 1K/2K/4K resolution.
Everything else has a default, which makes it the quickest tier to throw an idea at.
Pick 4.0 when you are still deciding. That covers the first several runs of almost every image anyone makes here: testing whether the composition works, whether the subject reads at a glance, whether the idea was any good. The prompt you arrive at transfers to 4.5 and 5.0 Pro unchanged, so nothing spent down here is wasted. Move up the moment either of the two known weaknesses matters — writing that has to be legible inside the frame, which is 4.5, or a human face doing something difficult, which is 5.0 Pro. And stay here permanently for decorative work: backgrounds, textures, and anything that will sit behind text at low opacity.
Roughly a fifth less than 4.5 and a third less than 5.0 Pro at its basic setting.
The resolution is a request field here rather than something implied by a quality switch.
Whatever wording works here works on 4.5 and 5.0 Pro without rewriting.
What the entry tier accepts, and where the tiers above it pick up.
One prompt in, one image out, with no source material required.
square, square_hd, portrait_4_3, portrait_3_2, portrait_16_9, and four landscape equivalents up to 21:9.
A real resolution field, which the 4.5 and 5.0 tiers replaced with a quality switch.
Reuse a seed to get closer to a previous run rather than starting over each time.
Taken from the published provider schema for this model, which is the API it runs on here.
Jobs where run count matters more than the last increment of fidelity.
The first several attempts at any image, before the prompt has settled.
Images that will sit behind type at low opacity and never be inspected.
One object, one light source, no writing on the label.
Anywhere you need many images rather than one good one.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, the safety check runs before the model does, and a failed run is refunded automatically. 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.
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 “SAYMAKER 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. Operator details, the full model list, and the refund window are on the about page.
Powered by
Use one balance across every image and video model on the shelf — the main ones are listed below — and check the credit cost before each request.
Video models
Image models
Model credit guide
A failed run is refunded automatically. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.5 | The long-take tier, priced per second: 5s 480p ≈ 525 credits, 5s 720p ≈ 1,185. A full 30s take runs ≈ 3,150 at 480p and ≈ 7,090 at 720p. |
| Video | Seedance 2.0 | Supplying a starting frame or clip costs less than starting from words alone: 6s 720p ≈ 565 credits from an image, ≈ 925 from text. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 6s 720p text-to-video ≈ 745 credits. |
| Video | Seedance 2 Mini | The cheapest tier of the Seedance 2 family. 6s 720p ≈ 465 credits, 6s 480p ≈ 215 credits. |
| Video | Seedance 1.5 Pro | Audio doubles the rate. 6s 720p ≈ 85 credits silent, ≈ 160 with audio; 1080p ≈ 175 and ≈ 340. |
| Video | Veo 3.1 | Billed per video, not per second, and the tier you pick is the whole price: Lite ≈ 115 credits at 720p and ≈ 135 at 1080p, Fast ≈ 225 at 720p, Quality ≈ 940 at 720p. |
| Video | Kling 3.0 | Audio raises the rate by about half. 6s 720p ≈ 320 credits silent, ≈ 455 with audio; 1080p ≈ 405 and ≈ 610. |
| Video | MiniMax H3 | Fixed 2K, no resolution ladder. Priced per second — a 5s clip ≈ 395 credits. |
| Image | Nano Banana 2 | Generate or edit from text and images. 20 credits per 1K image, 30 at 2K, 45 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. 30 credits per 1K or 2K image, 55 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant, and the everyday editing price: a flat 15 credits per image at every resolution. |
| Image | GPT Image 2 | The lowest-cost premium image model here. 10 credits per 1K image, 15 at 2K, 30 at 4K. |
| Image | GPT Image 2.5 Flare | OpenAI's newest image model, on the fast tier. 25 credits per 1K image, 40 at 2K, 60 at 4K — same price whether you generate or edit. |
| Image | GPT Image 2.5 Sunburst | The 2.5 tier for edits that touch only what you named. Same price as Flare: 25 credits per 1K image, 40 at 2K, 60 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — 20 credits per image whether you generate or edit. |
Common questions about the entry tier of the Seedream family.
4.0 takes a named framing preset and a 1K/2K/4K resolution; 4.5 takes an aspect ratio and a basic/high quality switch. The practical difference is text — 4.5 renders writing inside an image far more reliably — and 4.5 costs a little more.
About 20 credits, and the resolution you pick does not change it. Seedream 4.5 is about 25 and Seedream 5.0 Pro about 30 at its basic setting.
Only once you like the image. Resolution does not change composition or subject, so generating your test runs at 1K and the keeper at 4K loses nothing.
That is this tier's known weakness. Legible text inside an image is what Seedream 4.5 improved; if the words matter, generate there instead.
Reusing a seed keeps a new run close to a previous one, so you can change one part of the prompt without the whole image rearranging itself.
Credits are returned automatically, unless the request broke the content policy.
For its price, no — 20 credits flat, any use. The 5.0 line renders finer detail, but 4.0's output is stable and its look is well understood, which matters when you are matching images made months ago.
Pick a framing preset, describe the subject and the light, and run it a few times.