Seedream 5 Lite
prospolabs/seedream-5-lite
ByteDance's reasoning-based image model with chain-of-thought before generation. Generate from text or compose up to 10 reference images per edit. $0.014 per image — the cheapest reasoning image model on the web.
Starts from
Retail $0.035
Input
Typically completes in 5–15s.
Result
IdleOr write your own prompt and hit Generate.
01overview
What is Seedream 5 Lite?
Seedream 5 Lite is ByteDance's reasoning-based image model: it runs a chain-of-thought pass before it draws, so prompts about camera angle, lighting, spatial layout and composition land more reliably than with one-shot generators. On Prospolabs it runs at $0.014 per image — 60% below the $0.035 retail rate and the cheapest model on the platform.
It does two jobs from one endpoint. In generate mode it goes text-to-image; in edit mode it composes up to 10 reference images into a single result for multi-subject scenes, style transfer and material swaps. Prospolabs handles sizing automatically, so there are no resolution or aspect-ratio knobs to tune — you just prompt.
Prospolabs is pay-as-you-go in plain USD: no tokens, no monthly seat, no credit bundles to forecast. Top up from $5 (about 357 images at this price), pay the same rate in the UI and over the API, and failed runs are refunded automatically.
$0.014 per image — 60% off, the cheapest reasoning model on Prospolabs
Retail for Seedream 5 Lite sits around $0.035/image; Prospolabs runs it at $0.014 — roughly 60% off and the lowest price of any model on the platform. That makes 100 images $1.40 and 1,000 images just $14, billed in plain USD with no subscription and no tokens. Start from a $5 top-up (about 357 images) and pay the identical rate whether you generate from the playground or the API.
02capabilities
What Seedream 5 Lite can do
Reasoning before generation
Seedream 5 Lite plans the image with a chain-of-thought step before rendering, which is why complex, multi-clause prompts hold together instead of collapsing into the first plausible composition. It is the cheapest reasoning image model on the web at $0.014/image on Prospolabs.
Text-to-image (generate mode)
Send a prompt and a num_images value (1-6) and get clean, production-usable frames back. Sizing is handled automatically, so a single field controls how many variations come back in one call — each image bills at the flat $0.014 rate.
Multi-reference editing (edit mode)
Pass up to 10 reference images alongside your instruction to merge subjects, transfer a style, swap materials or change a scene while keeping characters consistent. Ideal for product shots and storyboards where the same subject must reappear across frames.
One flat USD price
Every image — generate or edit, UI or API — costs the same $0.014 on Prospolabs. No size-based tiers, no surcharge for editing, and no separate add-on fees to track. What you see is what each generation costs.
03use cases
What people build with Seedream 5 Lite
High-volume e-commerce imagery
Generate or re-style catalog and product shots at scale. A 100-image batch is $1.40 and 1,000 images is $14 on Prospolabs — cheap enough to regenerate whole catalogs without budget anxiety.
Consistent characters and storyboards
Use edit mode with up to 10 reference images to keep a character or product looking the same across a sequence of frames, comics or campaign variations.
Marketing and social asset sets
Spin coordinated banner, ad and social variants from one prompt. The reasoning pass keeps layout and lighting on-brief, so fewer regenerations are needed per usable asset.
Prototyping and drafts at the price floor
Because it is the cheapest model on Prospolabs, Seedream 5 Lite is the natural default for concepting, thumbnails and iteration — reserve pricier models for the few hero frames that truly need them.
04api
Call Seedream 5 Lite from your code
The Prospolabs API is one POST endpoint. Authenticate, pick the mode, and read back an output URL. The API price matches the UI exactly, and any failed run is refunded automatically.
Full Seedream 5 Lite API reference →- 01
POST to /v1/generate with model "seedream-5-lite", mode "generate", your prompt, and num_images (1-6).
- 02
For editing, set mode "edit" and attach up to 10 reference image URLs to compose multi-subject or style-transfer results.
- 03
Each call returns a generation_id and an output_url that stays valid for 7 days — download or rehost before it expires.
- 04
Bill per image at $0.014: a num_images batch simply multiplies that rate, so 100 images is $1.40. Top up from $5 to get started.
?questions
Seedream 5 Lite FAQ
$0.014 per image, flat — about 60% below the typical $0.035 retail rate and the cheapest model on Prospolabs. It is the same price in generate and edit mode, in the UI and over the API.
It runs a chain-of-thought reasoning pass before drawing, so detailed prompts about composition, lighting and spatial layout are followed more faithfully. On Prospolabs it is the cheapest reasoning image model on the web.
Pricing is linear at $0.014/image, so 100 images is $1.40 and 1,000 images is $14 on Prospolabs. You can request up to 6 images per call with the num_images parameter.
Yes. Set mode to "edit" and pass up to 10 reference images to compose multi-subject scenes, transfer styles, swap materials or keep a character consistent across frames.
Not directly — Prospolabs uses automatic image sizing for Seedream 5 Lite, so there are no separate resolution or aspect-ratio controls. You set num_images (1-6) and the prompt does the rest.
Pay-as-you-go in USD — no subscription, no tokens. Top up from $5 (around 357 images at this price) and only pay for what you generate. Failed runs are refunded automatically.
Yes. Prospolabs charges the same $0.014/image whether you generate in the playground or POST to /v1/generate. Each API response returns a generation_id and an output_url valid for 7 days.
It is the cheapest image model on Prospolabs, which makes it the default for high-volume and draft work. For different strengths you can also run Nano Banana 2, Nano Banana Pro or GPT Image 2 from the same account and balance.
related