Seedream 5.0 Lite is ByteDance Doubao's lean image generation model. Render at 2K or 3K, send reference images for image-to-image, and batch up to four variations per call — all through reApi's unified endpoint.
Free text · required · detailed prompts produce better results
Nine ratios from 1:1 to 21:9 — auto matches your reference image
2K for fast iteration · 3K for hero / print work
1–4 images per request · n>1 auto-enables group mode
Public HTTPS URLs · jpeg or png · ≤ 10 MB each
jpeg for size · png for transparency
disabled · auto — upstream auto-promotes to auto when count > 1
Only meaningful when sequential mode is auto
Off by default
Try one of these prompts
Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.
One sample at the cheapest tier (1 image).
Seedream 5.0 Lite is ByteDance's lean image model, exposed on reAPI as one async endpoint covering text-to-image and image-to-image through reference URLs. Each request picks 2K or 3K resolution and one of nine aspect ratio options including auto, and returns 1 to 4 images as JPEG or PNG — with sequential generation available for batched sets — billed per image.
Real-world workflows and production use cases you can build and ship with this model.
Seedream 5.0 Lite renders at 2K or 3K — sharp enough for the homepage hero, the product card, the campaign banner. Describe the shot, set resolution to 3K when the asset goes to print or a retina display, and the file lands in seconds. The same prompt template that drafted at 2K scales to 3K without recomposition, so a Seedream 5.0 Lite hero shot stays visually consistent across deployment surfaces.
Render at 3K
Set n to 4 and Seedream 5.0 Lite returns four variations in one request. The upstream auto-enables group mode under the hood, so you spend one API call and get a row of options to A/B in product reviews — no scripting an outer loop, no extra polling, one task ID covers the batch. Useful for moodboards, product card variants, and creative review.
Pass references in image_urls — product cutouts, environment plates, mood frames — and Seedream 5.0 Lite fuses them into one composition. All references are public HTTPS URLs (upload to your own R2 / S3 / OSS first). At 3K the texture detail holds up under print zoom; at 2K the file is light enough for ad-network creative pipelines.
Seedream 5.0 Lite covers a sharp two-tier resolution ladder. Send doubao-seedream-5-0-lite to /api/v1/images/generations with the resolution field set to 2K or 3K — same JSON shape across both. No client-side branching, no separate model id for higher detail.
1 credit equals $0.001 USD. Seedream 5.0 Lite charges a flat per-image rate at 2K and 3K, so estimating cost is a single-multiply on the n parameter. The estimator on this page costs your exact configuration before submit, and the Seedream 5.0 Lite charge in the dashboard equals what you saw at request time.
Submit a Seedream 5.0 Lite task, get a task_id back, poll until done. Failed and moderated requests are not charged. Drop Seedream 5.0 Lite into the same task plumbing you use for every other reApi model — the lifecycle is identical, the JSON shape matches the OpenAI image-generations contract.
Sign up at reApi and grab a key. Free signup credits cover dozens of Seedream 5.0 Lite generations, enough to ship the first 2K hero or campaign asset before topping up.
OpenPOST to /api/v1/images/generations with model: doubao-seedream-5-0-lite, a prompt, and optional size / resolution / n. Seedream 5.0 Lite returns a task_id immediately — your handler is free to move on.
OpenGET /api/v1/tasks/:id until status is completed. The Seedream 5.0 Lite response carries image URLs valid for 72 hours; mirror to your own storage if you need them longer.
Opencurl -X POST https://reapi.ai/api/v1/images/generations \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "doubao-seedream-5-0-lite",
"prompt": "a kitten yawning into the camera, cinematic warm tones",
"size": "16:9",
"resolution": "2K"
}'Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.
401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.
Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.
A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.
Common questions about this model.
Explore more models in the same category.
ByteDance
From $0.031 per image
Alibaba
From $0.031 per image
From $0.014 per image
SpaceXAI
From $0.016 per image
Try it in the playground or grab an API key to integrate now.