# Seedream 4.5 in Twistavo AI Studio

> Seedream 4.5 is an image generation and editing model from ByteDance Seed; on its internal MagicBench benchmark, ByteDance Seed reports significant gains over Seedream 4.0 in prompt adherence, alignment and aesthetics. Twistavo AI Studio offers Seedream 4.5. The cheapest run costs 9 credits: Seedream 4.5, quoted before you confirm.

Canonical page: https://twistavo.com/models/seedream-4-5
Updated: 2026-11-20

## Versions compared

Versions in Twistavo AI Studio, credits per run as quoted today

| Version | Type | Resolutions | Frame shapes | Inputs | Price per run | What the maker says, and what Twistavo offers |
| --- | --- | --- | --- | --- | --- | --- |
| Seedream 4.5 | Image | Default | 1:1, 4:3, 16:9, 3:2, 21:9, 3:4, 9:16, 2:3 | Prompt + reference images (up to 14) | 9 credits | ByteDance Seed: designer-level composition and typography with clear and readable small text, ideal for posters, brand visuals and other creative scenarios; it identifies the main subjects in multi-image editing and strictly preserves the details of the reference images. ([source](https://seed.bytedance.com/en/seedream4_5)) On Twistavo: inputs: prompt, reference images |

- [Seedream 5.0](https://twistavo.com/models/seedream-5-0): for the Lite, Pro and Flash versions of the next generation

## Limits and known issues

- Through BytePlus ModelArk, Seedream 4.5 takes up to 14 reference images, and the input and output images of one request add up to 15 at most (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/image-generation-api))
- BytePlus offers Seedream 4.5 output at 2K or 4K through its API; AI Studio has no size setting for it and offers a quality setting instead. ([source](https://docs.byteplus.com/en/docs/ModelArk/image-generation-api))
- Reference images must be jpeg, png, webp, bmp, tiff, gif, heic or heif, up to 30 MB each, with a width-to-height ratio between 1/16 and 16 (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/image-generation-api))
- Through the API, Seedream 4.5 can also return a series of related images for storyboards, comics or emoji packs, triggered by words such as a series or a set (BytePlus); the table above shows what AI Studio offers. ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))

## How to get good results

1. Describe subject, action and setting in plain sentences, then the style, colour, light or composition: A girl in a lavish dress walking under a parasol along a tree-lined path, in the style of a Monet oil painting, rather than Girl, umbrella, tree-lined street, oil painting texture (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))
2. Short, precise prompts usually beat piles of ornate words: BytePlus says Seedream 4.5 needs less description than Seedream 3.0. ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))
3. State what the image is for and its type, such as Design a logo for a gaming company, and put every word that must appear in double quotation marks (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))
4. For an edit, name the target without vague pronouns and say what stays unchanged: Dress the tallest panda in pink Peking Opera costume and headgear, keeping its pose unchanged, not Put that one in pink clothes (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))
5. When a region is hard to describe, draw an arrow, a bounding box or a doodle on the image to mark what to change (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))
6. With several images, say what comes from each one, such as replace the character in Image 2 with the character from Image 1, in the style of Image 3 (BytePlus). ([source](https://docs.byteplus.com/en/docs/ModelArk/seedream-4-0-5-0-prompt-guide))

## Questions

### Is this the same model as the newer Seedream versions?

No. It is an earlier generation with its own page; the newer Seedream versions are listed separately in the model list.

### Can I translate the text in an existing image?

ByteDance Seed shows it rewriting the text of an image in another language while keeping the rest. Add the image as a reference and name the target language.

### Which inputs and shapes are offered?

A prompt, reference images, and frame shapes including 1:1, 4:3, 16:9, 3:2, 21:9, 3:4, 9:16 and 2:3.

### Does the quality setting change the price?

The quote in the composer is the reference: change the setting and the quote updates before you confirm.

### How much does Seedream 4.5 cost per image?

Price per image in Twistavo AI Studio: Seedream 4.5: 9 credits. The composer also has a Quality setting (Basic or High), and the quote shows the price before you confirm.

### Is Seedream 4.5 free?

Seedream 4.5 runs are paid in credits in Twistavo AI Studio. 1 free image after signup, no card (GPT Image 2.6, 1k). Videos need a plan. Plans from $3.99/month. Credits never expire and every generation is quoted before it starts.

### Who makes Seedream?

ByteDance Seed makes the Seedream image models: its site lists them next to its Seedance video models, and BytePlus serves them to developers through its ModelArk API. Twistavo is an independent service and is not affiliated with ByteDance. ([source](https://seed.bytedance.com/en/seedream4_5))

### What is Seedream 4.5 good for?

ByteDance Seed presents it for professional visuals with high consistency: edits that keep a reference image's facial features, lighting and colour tone, multi-image edits, and posters, logos and brand visuals with clear small text. Its own examples include an e-commerce product image with a title and a price line, and translating the text of an image into another language. ([source](https://seed.bytedance.com/en/seedream4_5))
