Skip to main content

Seedream 4.5

Seedream 4.5 is the model ByteDance's Seed team released on 23 December 2025, the last of the 4.x line before 5.0 arrived in July 2026. It unified generation and editing in one architecture, outputs at 2K and 4K, and is still sold at a flat $0.04 per image on fal.

Anvisha PaiAnvisha Pai, Co-founder & CEO, ModaTested

ByteDance · released

Seedream 4.5 self-portrait, editorial fashion photograph
Seedream 4.5 drawing itself as editorial fashion photograph, generated by us.
What we asked for

A self-portrait of Seedream 4.5, ByteDance's image model, as an editorial fashion photograph: a model in a sculptural white paper dress against a saturated teal seamless, hard studio flash, a subtle '4.5' embossed on a paper clutch, magazine cover crop.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":1536,"height":2048}, num_images: 1
returned 1920 by 2560 pixels

At a glance

Price
A flat $0.04 per image.
Sizes
2K or 4K; every sample below is 2K.
Reference images
Up to ten on the edit endpoint.
Superseded
Seedream 5.0 Pro and Lite are the current models; 4.5 stays on the API.
One model for both
Generation and editing in a single architecture.

Other Seedream models: Seedream 5.0 Pro, Seedream 5.0 Lite. Compare all 3.

Anvisha's take

4.5 is the one I would retire. The poster lost its layout, with 'SESSIONS' floating below the date and a stray letter, the self-portrait put a near-copy of a real fashion magazine's masthead across the top, and the edit replaced the dripper with a different design. It is quick and cheap, and the photographic samples are decent, but Seedream 5.0 Lite costs less and 5.0 Pro does everything else better.

Seedream 4.5 real output: the same 12 prompts every model gets

Every model on this site gets the same 12 prompts, so you can compare like with like. These are the pictures Seedream 4.5 returned, one try each, nothing re-rolled or retouched. Generated on . The 12 pictures took 9 to 16 seconds each and cost $0.48 in total.

Seedream 4.5 output for the text in an image prompt
Text in an image16s · $0.04

Spelling, kerning and hierarchy of three lines of set text.

What we asked for

A concert poster that reads 'MODA SUMMER SESSIONS' in large bold type, with the subtitle 'Friday 12 June, 8pm, Roundhouse' and a smaller line 'Tickets at the door'. Risograph style, two colours, tangerine and navy, grainy ink texture.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":1536,"height":2048}, num_images: 1
returned 1920 by 2560 pixels

Seedream 4.5 output for the product on white prompt
Product on white9s · $0.04

Catalogue lighting, a clean seamless background and a believable contact shadow.

What we asked for

Studio product photograph of a matte black ceramic pour-over coffee dripper on a seamless white background, soft top light, subtle contact shadow, centred, catalogue style.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":2048}, num_images: 1
returned 2048 by 2048 pixels

Seedream 4.5 output for the face and hands prompt
Face and hands10s · $0.04

A natural face, correct finger count and hands holding an object.

What we asked for

Close-up portrait of a woman in her 40s with short grey hair, laughing, holding a red ceramic mug with both hands, natural window light, 85mm lens, shallow depth of field.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":1536,"height":2048}, num_images: 1
returned 1920 by 2560 pixels

Seedream 4.5 output for the logo mark prompt
Logo mark9s · $0.04

A flat single-colour mark with a legible wordmark and no gradients or photo texture.

What we asked for

A flat vector logo mark for a bicycle repair shop called 'Spoke & Chain': a single-colour icon combining a chain link and a wheel spoke, with the wordmark below in a geometric sans-serif, black on white.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":2048}, num_images: 1
returned 2048 by 2048 pixels

Seedream 4.5 output for the busy scene prompt
Busy scene11s · $0.04

Many distinct objects and people in one frame without merging or duplicating them.

What we asked for

A crowded Saturday farmers market seen from slightly above: at least twelve stalls with striped awnings, shoppers carrying tote bags, a dog on a leash, crates of oranges, a musician with a guitar, bunting overhead, late morning light.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":1536}, num_images: 1
returned 2560 by 1920 pixels

Seedream 4.5 output for the specific art style prompt
Specific art style11s · $0.04

A named historical style applied faithfully rather than a generic painterly look.

What we asked for

A lighthouse on a rocky cliff during a storm, in the style of a Japanese ukiyo-e woodblock print, with visible wood grain texture, flat colour areas and a bokashi gradient sky.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":1536,"height":2048}, num_images: 1
returned 1920 by 2560 pixels

Seedream 4.5 output for the chart prompt
Chart11s · $0.04

Six labelled bars at the stated heights, in order, with readable axes.

What we asked for

A clean bar chart titled 'Monthly active users' showing six months, January through June, with the values 120, 150, 180, 170, 220 and 260, labelled axes, white background, one accent colour, no 3D effects.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":1536}, num_images: 1
returned 2560 by 1920 pixels

Seedream 4.5 output for the wide aspect ratio prompt
Wide aspect ratio11s · $0.04

A 21:9 composition that leaves the left third clear for headline text.

What we asked for

Website hero banner: a wide panorama of a sunlit modern co-working space with plants, the left third of the frame kept as clear wall space for text, warm neutral palette.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2688,"height":1152}, num_images: 1
returned 4480 by 1920 pixels

Seedream 4.5 output for the infographic with steps prompt
Infographic with steps10s · $0.04

Four numbered steps in the given order with their exact wording, plus an icon each.

What we asked for

An infographic card titled 'How to brew pour-over' with four numbered steps in this order: 1. Rinse the filter, 2. Bloom for 30 seconds, 3. Pour in slow circles, 4. Wait 3 minutes. A simple icon beside each step, muted palette, clean sans-serif type.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":1536,"height":2048}, num_images: 1
returned 1920 by 2560 pixels

Seedream 4.5 output for the food photography prompt
Food photography9s · $0.04

Appetising, menu-grade food with correct textures and no melted anatomy.

What we asked for

Overhead photograph of a margherita pizza on a wooden board, fresh basil leaves, one slice pulled out with the cheese stretching, restaurant menu photography, natural light.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":2048}, num_images: 1
returned 2048 by 2048 pixels

Seedream 4.5 output for the hard composition prompt
Hard composition10s · $0.04

Exact counts and left-right placement of five named objects.

What we asked for

A glass of water on a wooden table with exactly three ice cubes in it, a yellow pencil resting across the rim of the glass, a blue book to the left of the glass and a red apple to the right, photographed from table height.

fal-ai/bytedance/seedream/v4.5/text-to-image
image_size: {"width":2048,"height":1536}, num_images: 1
returned 2560 by 1920 pixels

Seedream 4.5 output for the edit an existing image prompt
Edit an existing image14s · $0.04

Change one object's colour and add a part while leaving background, light and framing untouched.

What we asked for

Change the coffee dripper to a deep forest green and add a small brass handle on its right side. Keep the background, lighting and composition exactly as they are. (applied to the product on white result above)

fal-ai/bytedance/seedream/v4.5/edit
image_size: {"width":2048,"height":2048}, num_images: 1
returned 2048 by 2048 pixels

How the cost under each picture was worked out: fal's flat $0.04 per image for Seedream 4.5 (fal model page, read 2026-09-09).

Seedream 4.5 pricing

WhereUnitPriceSource
falImage$0.04fal model page, read

Is Seedream 4.5 free?

ByteDance's consumer apps, Dreamina (CapCut's creative app) and Doubao in China, generate with Seedream and give free daily credits. The API, through Volcano Engine, BytePlus and hosts like fal, is paid per image.

  • Dreamina: free daily credits for images and video, Seedream 5.0 Pro as the flagship model and 5.0 Lite as the fast option.
  • Doubao and Jimeng: ByteDance's China apps, free with limits.
  • API and hosts: pay per image.

Dreamina, read

Where to use Seedream 4.5

  • Dreamina · ByteDance's consumer app, free daily credits.
  • BytePlus ModelArk · ByteDance's API for international customers.
  • fal · Per-image pricing, generate and edit endpoints.
  • Moda · Available inside the Moda design agent as Seedream 4.5.

Generate with Seedream 4.5 in Moda

Pick Seedream 4.5 in the Moda design agent, then edit the result as vectors, text and layers rather than a flat image.

Open Moda →

What we noticed running Seedream 4.5

  • Layout failure: 'MODA SUMMER' and the date are set fine, then a lone 'D' and 'SESSIONS' appear below, out of order. The only broken poster in this batch. See the text in an image picture.
  • Trademark leak: asked for an editorial fashion photograph, the self-portrait added a masthead imitating a famous fashion magazine, misspelled, plus a barcode. Check anything editorial for borrowed brands.
  • Edits are re-draws: the recolour arrived on a rounded, fluted dripper with a black base and a stub where the handle should be. See the edit an existing image picture.
  • Pencil across the rim, book left, apple right, four or five ice cubes. See the hard composition picture.
  • Quick: 9 to 16 seconds per picture at a flat $0.04, every output 2K or larger because of the endpoint's minimum size.

Seedream 4.5 limits and API parameters

LimitValue
Output size1920 to 4096 pixels a side, or at least 2560 by 1440 in total.
fal API reference, read
Reference imagesUp to 10 per edit.
fal API reference, read
Every setting, for developers

fal (fal-ai/bytedance/seedream/v4.5/text-to-image)

fal API reference, read

ParameterTypeValuesDefault
promptstringAny text
Required.
image_sizeenum or objectsquare_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9, auto_2K, auto_4K, or {width, height} from 1920 to 4096auto_2K
num_imagesinteger1 or more1
max_imagesinteger1 or more1
seedintegerAny integer
enable_safety_checkerbooleantrue, falsetrue
image_urlslist of URLsUp to 10
Edit endpoint only.

What Seedream 4.5 will not draw

ByteDance's models run behind fal's safety checker, on by default and only switchable off with account authorisation, and under fal's partner terms for commercial use. ByteDance publishes its own content rules for Dreamina and the BytePlus API; the practical boundaries are sexual content, minors, violence and deceptive depictions of real people. Outputs from the consumer apps carry an AI-generated label.

fal API reference, read

Alternatives to Seedream 4.5

Frequently asked questions

What is Seedream 4.5?

Seedream 4.5 is ByteDance's December 2025 image model, unifying generation and editing, with 2K and 4K output at $0.04 per image on fal. Seedream 5.0 replaced it in July 2026.

Is Seedream 4.5 still worth using?

Only if a pipeline is tuned to it. Seedream 5.0 Lite is cheaper and 5.0 Pro is better on text; compare the samples.

How this page is made

Prices, limits and settings are read from the linked vendor and host pages on the dates shown. Every picture was generated by us through fal with the request shown under it, and the original files are kept unedited. A generation that failed is shown as a failure. Model and vendor names are used to identify the products; no vendor imagery appears on this page.

Anvisha Pai

Anvisha Pai

Co-founder & CEO, Moda

Anvisha is the CEO of Moda and a repeat, Y Combinator-backed startup founder. She was previously a PM at Dropbox. She believes nobody should need a design degree to make something that looks great.