How long does a Seedream 5.0 Flash text-to-image render take?

In our runs most 2K renders finished in 29 to 37 seconds, one took 92, and a 1.5K render came back in about 21. The task is asynchronous, so poll for a terminal state instead of holding the request open. We have not timed Seedream 5.0 Pro under matched conditions, so we make no speed comparison between the two.

What pixel size does each Seedream 5.0 Flash tier produce?

It depends on the ratio as well as the tier. Measured outputs: 2K at 21:9 came back at 3136 by 1344, 2K at 4:3 at 2368 by 1776, 1.5K at 4:5 at 1376 by 1712 and 1.5K at 5:4 at 1712 by 1376. Treat the tier as a pixel budget spread over the chosen shape, and measure the file when a layout depends on an exact figure.

Why did my Seedream 5.0 Flash image come back square?

Because aspect_ratio was left out. The text-to-image endpoint has no automatic ratio, so an omitted field renders 1:1. Name one of the fifteen shapes on every call; the Flash editing endpoint is different and defaults to the shape of the first reference.

How should I write a Seedream 5.0 Flash prompt that includes text?

Quote each line exactly, say where it sits and how large it is relative to the rest, and finish with an instruction to use only the supplied text. That is how the LATE SHIFT poster brief was written, and all three lines came back as written. Short lines survive best; proofread small print before publishing.

Can I change just the headline after the image is made?

Yes, without regenerating. Send the finished image to Seedream 5.0 Flash Layer Decomposition, which returns a repaired base plate plus a transparent PNG for each element; replace the title layer in your design tool and place it back. Every layer carries a name and a position box, so the pieces snap back into place.

Does a longer prompt or png output cost more?

No. Seedream 5.0 Flash text-to-image bills per image, and the price does not change with prompt length, aspect ratio, tier or output_format. png is lossless and the safer handoff into retouching; jpeg is smaller for direct delivery.

Can one call return several variations?

No. Each call returns one image, and the schema has no batch or output-count field. Loop separate calls for variations; each queues and bills on its own, and because there is no seed they differ in more than the clause you changed.

Is there a negative prompt field?

No. Seedream 5.0 Flash text-to-image has no negative-prompt field and rejects fields it does not declare. Describe what should be in the frame instead: asking for a plain black background works where asking for no clutter tends not to.

Seedream 5.0 Flash API — Uncensored NSFW Images & Layers

bytedance/seedream-5.0-flash/text-to-image

Seedream 5.0 Flash text-to-image is ByteDance's lighter Seedream tier, turning a prompt of up to 5,000 characters into one image. Choose from fifteen aspect ratios between 1:3 and 3:1 at 1K, 1.5K or 2K, all at one price.

From $0.029/image

Run model View API

Input

Text to image

Prompt346/5000

Aspect ratio

Resolution

Output format

JPEGPNG

Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run

Example result

Animate

Edit

Split into layers

{
  "model": "bytedance/seedream-5.0-flash/text-to-image",
  "input": {
    "aspect_ratio": "9:16",
    "resolution": "2k",
    "output_format": "png",
    "prompt": "A woman in her thirties in a deep emerald velvet slip dress leans against a rain-streaked window at night, one thin strap slipping off her shoulder, city neon smeared in pink and teal across the glass and her collarbone, she looks back over her shoulder at the camera with a slow half-smile, 35mm film look, shallow depth of field, photorealistic"
  }
}

Introduction

Seedream 5.0 Flash Text-to-Image creates a new image from a written brief. Use it for posters with real lettering, portraits, product and fragrance stills, and editorial or after-hours scenes when you want to start from words rather than a photo.

Developed by ByteDance (model ID dola-seedream-5-0-flash-260915). Every resolution tier costs the same per image, so the default is the sharpest one.

Choose a Seedream 5.0 Flash mode by starting material

Starting material	Mode
A written brief for a new picture	Seedream 5.0 Flash Text to Image
One to ten images to revise or combine	Seedream 5.0 Flash Image Editing
A finished image whose parts you need separately	Seedream 5.0 Flash Layer Decomposition

Key features & highlights

Lettering that comes out as written

The retro film poster on this page carries a cream serif title, a tagline and a ticket stub, all spelled as briefed. The perfume ad renders an accented headline correctly. Quote the exact copy and say where it sits.

Fifteen ratios and three flat-priced tiers

aspect_ratio runs from 1:3 to 3:1 and resolution is 1k, 1.5k or 2k (default 2k). In our runs 2K at 21:9 came back at 3136 by 1344 and 1.5K at 4:5 at 1376 by 1712.

Unfiltered imagery without a switch

Unfiltered prompts run on the same endpoint as everything else, with no maturity flag.

5,000-character briefs

Enough for subject, wardrobe, light, lens and grade in a single prompt.

Request parameters

Request parameters for Seedream 5.0 Flash · Text to Image
Parameter	Required	Description
`prompt`	Yes	Up to 5,000 characters. Spend it on who is in frame, what they wear, where the light comes from and the lens feel. at least 1 characters · at most 5000 characters
`aspect_ratio`	No	1:1 by default, or one of fourteen other shapes from 1:3 to 3:1, including 4:5, 21:9 and 9:21. `1:1` · `4:3` · `3:4` · `16:9` · `9:16` · `3:2` · `2:3` · `4:5` · `5:4` · `21:9` · `9:21` · `2:1` · `1:2` · `3:1` · `1:3` · Default: `1:1`
`resolution`	No	1k, 1.5k or 2k; 2k by default. All three tiers cost the same per image, so the default is the sharpest. `1k` · `1.5k` · `2k` · Default: `2k`
`output_format`	No	jpeg by default, or png. png is lossless and the safer handoff into retouching or compositing; it changes the delivered file, not the price. `jpeg` · `png` · Default: `jpeg`

Pricing

$0.029 / image

Use cases

Film posters and ad stills

Put the title and tagline in quotation marks and describe the composition around them.

Portrait series for a character

Write one strong treatment, then carry the face into new scenes with the editing endpoint.

Widescreen banners

Set 21:9 for a wide, cinematic frame such as the poolside swimmer on this page.

After-hours editorial

Soft morning window light, silk on a Paris balcony, two dancers alone in a ballroom.

Explore key use-case tutorials

Prompt guide & pro tips

Write the shot

Say who or what is in frame, what they wear, where the light comes from and the lens or film feel. The rainy-window portrait on this page used a 35mm film look and city neon in pink and teal.

Choose the shape with aspect_ratio.
Describe subject, wardrobe and setting.
Name the light: direction, color and hardness.
Quote any text exactly and say where it sits.

Retro film poster, portrait 3:4. A woman in a red trench coat holds a paper coffee cup in front of a rainy-night diner cart under a red neon sign.
Cream serif title across the top: "LATE SHIFT". Below it, a smaller line: "Some orders never get cancelled".
Yellow ticket stub in the lower right corner reading "ADMIT ONE".
Use only the supplied text, with no extra badges or logos.

Draft at 1k, deliver at 2k

All tiers cost the same, so there is no saving in drafting low, but 1k is quicker to review when you are iterating on a composition.

Content boundaries

Seedream 5.0 Flash · Text-to-Image — Uncensored. SpicyAPI adds no platform-side content filter. Model restrictions still apply, and results are not guaranteed. You are responsible for complying with our content policy.

API quick start

# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $SPICY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
  -d '{
  "model": "bytedance/seedream-5.0-flash/text-to-image",
  "input": {
    "prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
    "resolution": "2k",
    "aspect_ratio": "1:1",
    "output_format": "jpeg"
  }
}'

# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $SPICY_API_KEY"

See the API reference for authentication, the complete schema, polling, and error handling.

Related models and next tasks

Missing the model you need?

Tell us what to add and where you would use it. Requests go to the people who decide what gets onboarded next.

Spelled-right posters and editorial portraits at one price per image

Seedream 5.0 Flash's text-to-image endpoint turns a written brief into one image in any of fifteen aspect ratios. The 1k, 1.5k and 2k tiers all cost the same, so the default is the sharpest. In our runs it printed a retro poster's title, tagline and ticket stub exactly as quoted, and it handled Mild-level fashion editorials with no flag and no separate endpoint.

Open the playground

Spelled-right posters and editorial portraits at one price per image

Aspect ratios: 15 options from 1:3 to 3:1; 1:1 if left unset
Resolution: 1k, 1.5k or 2k (default 2k), same price at every tier
Prompt length: 1 to 5,000 characters
Output format: jpeg (default) or png
Seed: None — identical requests return different frames
Content policy: No content filter, no maturity flag

Flat tiers: The tier changes the pixels, not the price1k, 1.5k and 2k bill the same per image, which is why 2k is the default. Drop a tier only for a smaller file or a quicker look: a 1.5K frame came back in about 21 seconds in our runs, most 2K frames in 29 to 37.
Lettering: Quoted copy comes back spelled as writtenA poster brief with a title, a tagline and a ticket stub returned all three correctly, and a fragrance ad kept the accent marks in its headline. Quote every line and say where it sits.
Next step: Any result can be split into layers afterwardsSend a finished frame to this family's Layer Decomposition endpoint for a clean base plate and one transparent PNG per element, so a headline can be rewritten without regenerating the picture. On the web it is one click from the result.

What the Seedream 5.0 Flash text-to-image API gives you

Posters and ads with the words spelled right

Put every line of copy in quotation marks and name its position: title across the top, tagline underneath, a stub in the lower right. Our retro diner poster came back with its serif title, tagline and ADMIT ONE stub spelled correctly, and the perfume ad kept the accents in NOIR ÉTÉ. Proofread before publishing all the same; one clean run is a good sign, not a guarantee.

Fifteen frames from the same brief

aspect_ratio runs from 1:3 to 3:1 and is read before the prompt, so the picture is composed for that shape rather than cropped into it. 9:16 suits a phone-first portrait, 4:5 a feed post, 21:9 a cinematic header; 2K at 21:9 came back at 3136 by 1344 in our runs. Leave the field out and the result is square.

Light and fabric carry the mood
The strongest frames in our tests were built from light: pink and teal neon across a rain-streaked window and a collarbone, morning sun through sheer curtains onto ivory lace, one chandelier beam over two dancers. Name the source, its colour and where it lands, then let texture words such as velvet, satin and linen do the rest.
Editorial work on the same call
Mild-level briefs are ordinary prompts here. SpicyAPI adds no content filter, and there is no maturity flag or second endpoint to look up; the model's own limits still apply. The rights to what you publish, and keeping it lawful, stay with you.
Priced below Seedream 5.0 Pro per image
A Flash image costs less than a Seedream 5.0 Pro image at either of Pro's tiers, and every Flash tier is the same price, which makes it a sensible place to iterate on a layout or a series. We have not run a matched quality test between the two, so compare them on your own brief before committing a batch.
No seed, so the brief is the record
Two identical requests return two different frames. Keep the exact wording, the ratio and the tier next to any image meant to anchor a series; replaying them returns a sibling, and this family's editing endpoint is the way to carry the same face into new scenes.

Where Seedream 5.0 Flash text-to-image fits

Product and fragrance
Ad stills with the brand line in place
Describe the object, the surface and the light, then quote the headline and the small print with their positions. The perfume test set smoked glass and a gold ribbon on wet black marble, with the name above the bottle and the product line beneath it.
Character work
A portrait strong enough to anchor a series
Write one detailed treatment of a face, a wardrobe and a light, generate until a frame lands, then take that frame to the editing endpoint to place the same person in new scenes. Describing the face from scratch every time drifts; a reference image does not.
Fashion and editorial
Styled scenes that read as photography
Two figures alone in a ballroom, a portrait at a window: keep the brief concrete about fabric, light and pose and the result reads like a styled shoot rather than stock imagery. No switch is needed for this material.
Headers and banners
Wide frames composed for the strip
Pick 21:9 or 3:1 when the image runs across a page header, and say where the empty space should fall. The poolside test fitted a lone swimmer, a towel and teal light ripples across a 21:9 frame without anything feeling squeezed.

How to call the Seedream 5.0 Flash text-to-image API

Name the ratio before writing
aspect_ratio defaults to 1:1, so choose the placement's shape first; the composition is built for it from the first word.
Write the shot, then the words
Subject, wardrobe, light direction and colour, lens or film feel, as plain clauses within 5,000 characters. Put any text in quotation marks with its position.
Keep 2k unless you need a smaller file
All three tiers cost the same, so 1k or 1.5k saves file size and a few seconds, not money.
Submit and poll for a terminal state
The task runs asynchronously; most 2K renders in our tests finished in 29 to 37 seconds and one took 92. Poll until it finishes, then fetch the image from platform storage.
Keep the brief with the keeper
The prompt, ratio and tier are the only record of how a frame was made. From there, split it into layers or send it to the editing endpoint.

Seedream 5.0 Flash text-to-image worked examples

A poster with three lines of real copy

Every line is quoted and placed, and the last sentence stops the model adding badges of its own. This is the brief behind the LATE SHIFT poster.

Retro film poster, portrait 3:4. A woman in a red trench coat holds a paper coffee cup in front of a rainy-night diner under a red neon sign. Cream serif title across the top: "LATE SHIFT". Below it, a smaller line: "Some orders never get cancelled". Yellow ticket stub in the lower right corner reading "ADMIT ONE". Use only the supplied text, with no extra badges or logos.

A portrait lit by the city

The light source, its colours and where they land carry the whole frame; aspect_ratio is 9:16 in the request.

A woman in her thirties in a deep emerald velvet slip dress leaning at a rain-streaked night window, one thin strap slipped off her shoulder, pink and teal city neon on the glass and her collarbone, glancing back at the camera with a half smile, 35mm film look

A 21:9 header with room to breathe

aspect_ratio is 21:9 and the subject sits on one side, so the wide frame keeps usable space. Our poolside test used a brief along these lines.

Cinematic widescreen photograph, a swimmer in his twenties sitting alone at the edge of an empty indoor pool at night, towel over his shoulders, underwater lights throwing teal ripples across the tiled wall, subject on the right third, quiet mood

Seedream 5.0 Flash text-to-image questions

Decide the shape first, because the frame is composed for it from the first word and an omitted ratio comes back square. Then write the shot: who is in frame, what they wear, where the light comes from and what colour it is. Quote any words exactly with their position. Keep the 2k default, since every tier costs the same, and store the brief with any frame you keep.

All Seedream models4

Seedream 5.0 FlashCurrentText to image · Image editing · +1$0.026per image
Seedream 5.0 ProText to image · Image editing · +1ByteDance official price: $0.045From $0.036per image
Seedream 5.0 LiteText to image · Image editing$0.035per image
Seedream 4.0Text to image · Image editing$0.03per image

Related comparisons

Seedream 5.0 Flash vs Seedream 5.0 Pro Seedream 5.0 Flash vs Qwen Image 2.1 Seedream 5.0 Flash vs Qwen Image Edit Spicy Seedream 5.0 Flash vs Qwen Image 3.0 Pro Seedream 5.0 Flash vs Qwen Image 3.0 Seedream 5.0 Flash vs Z-Image Spicy Pro

Workbench