New · Image Generation

Every top image model, one API

Gemini, GPT Image, Seedream and more — call any image model through one unified API. One key, no queue.

Get API key →Browse all models

Webofox.design · type it, get the image→

OpenAI-compatible · one key for every model · pay as you go

Top image-model providers, unified

GoogleOpenAIVolcengineAlibaba BailianMicrosoft

Image models

Providers

Unified API

Prompt

Gemini 3 Pro Image · ~4s

Gemini 3 Pro Image⚡ ~4s

Five latest models · Preset demo — open ofox.design for real creationOpen ofox.design to create your own →

Model Lineup

Full Model Lineup

Covering Google, OpenAI, Volcengine, Alibaba Bailian, and Microsoft — switch in one place, compare on the same prompt, and always pick the perfect shot.

NEW

Qwen-Image 3.0 Pro

bailian/qwen-image-3.0-pro

Pin-sharp text down to 10px — top pick for posters and typography

Flagship$0.04/image

NEW

Qwen-Image 3.0

bailian/qwen-image-3.0

Same generation at a lower price — steady text and detail for daily work

Value$0.03/image

NEW

MAI Image 2.5 Pro

microsoft/mai-image-2.5-pro

Rivals GPT Image 2 on quality, with strong consistency for complex scenes

Flagship$0.11/image

NEW

MAI Image 2.5

microsoft/mai-image-2.5

Precise surgical edits — 5–10× faster at about 1/5 the cost

Value$0.048/image

NEW

MAI Image 2.5 Flash

microsoft/mai-image-2.5-flash

The fastest, most affordable model in the MAI lineup — built for high-volume batches

Turbo$0.027/image

GPT Image 2

openai/gpt-image-2

Balanced quality with stable, controllable text and layout

FlagshipImage output $30/M

GPT Image 1.5

openai/gpt-image-1.5

A stable classic — the all-purpose pick for any subject

ClassicImage output $32/M

Gemini 3 Pro Image

google/gemini-3-pro-image

Top-tier detail — the go-to for complex scenes and lighting

FlagshipImage output $120/M

Model	Max resolution	Text-to-image	Image-to-image	Price ⇅
OpenAI
GPT Image 2openai/gpt-image-2	4K3840	✓	✓	Image output $30/M
GPT Image 1.5openai/gpt-image-1.5	1.5K1536	✓	✓	Image output $32/M
Google
Gemini 3 Pro Imagegoogle/gemini-3-pro-image	4K4096	✓	✓	Image output $120/M
Gemini 3.1 Flash Imagegoogle/gemini-3.1-flash-image	4K4096	✓	✓	Image output $60/M
Gemini 3.1 Flash-Lite Imagegoogle/gemini-3.1-flash-lite-image	1K1024	✓	✓	Image output $30/M
Gemini 2.5 Flash Imagegoogle/gemini-2.5-flash-image	1K1024	✓	✓	Image output $30/M
Volcengine
Doubao Seedream 5.0 Provolcengine/doubao-seedream-5.0-pro	2K~4.6MP	✓	✓	$0.05/image
Doubao Seedream 5.0 Litevolcengine/doubao-seedream-5.0-lite	4K4096	✓	✓	$0.035/image
Doubao Seedream 4.5volcengine/doubao-seedream-4.5	4K4096	✓	✓	$0.04/image
Alibaba Bailian
Qwen-Image 3.0 Probailian/qwen-image-3.0-pro	2K~6.55MP	✓	✓	$0.04/image
Qwen-Image 3.0bailian/qwen-image-3.0	2K~6.55MP	✓	✓	$0.03/image
Microsoft
MAI Image 2.5 Promicrosoft/mai-image-2.5-pro	1K1024	✓	✓	$0.11/image
MAI Image 2.5microsoft/mai-image-2.5	1K1024	✓	✓	$0.048/image
MAI Image 2.5 Flashmicrosoft/mai-image-2.5-flash	1K1024	✓	✓	$0.027/image

Prices are from the ofox model marketplace (image-output basis); Volcengine, Bailian, and Microsoft bill per image while OpenAI/Google bill per token ($/million tokens). Because the units differ, price sorting is performed within each provider's group.

Why Ofox

Why Ofox for image generation

One prompt, the creative power of every major model — from spark to finished image, we handle everything in between.

sk-of-••••••••

Top models, one switch

Gemini, GPT Image, Seedream, Qwen-Image, and MAI Image all ready to go. Compare the same prompt side by side, keep whichever wins, and stop copy-pasting across platforms.

14 models5 providersSame-prompt comparison

Endless prompts, one-tap recreate

Thousands of curated prompts, filtered by theme and model — drop one in, tweak a few words, and generate. This inspiration library works out of the box in the ofox.design web app, no prompt magic required.

ThousandsOne-tap recreateBeginner-friendly

Try ofox.design →

→

Reference → image editing

Reference image · image-to-image

Pass in a reference image and it routes to the image editing endpoint (image-to-image) — style continuation, inpainting, and multi-image fusion, natively supported by gpt-image, Gemini, and Seedream.

Image-to-imageInpaintingMulti-image reference

Standard2K4K

16:9

1:1

9:16

Full-size 4K output

Standard / 2K / 4K resolution tiers plus multiple aspect ratios, with batch results in a single run — each model's native max resolution, delivered in one step.

Standard/2K/4KMultiple aspect ratiosBatch results

Web · No-code

ofox.design

Whatever you imagine, just draw it — one sentence, top models render it instantly.

ofox.design is a no-code creative image studio built on the same top models: type a sentence in plain language to get an image instantly, with thousands of built-in prompts for one-tap remixes, support for reference images / image-to-image, and full-size output from standard up to 4K. Your work is auto-archived and ready to recreate anytime. No API, no deployment — just open the page and go.

Open ofox.design →

New users get free creation credits on sign-up · No credit card required

ofox.design · Web

Quickstart

Generate images in a few lines of code

Call image models with your key — with the OpenAI-compatible endpoint, swap the model to switch vendors, or go straight to each provider's native API.

# 一个 OpenAI 兼容端点；换 model 即切换厂商
curl https://api.ofox.ai/v1/images/generations \
  -H "Authorization: Bearer $OFOX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3-pro-image",
    "prompt": "赛博朋克东京雨夜，霓虹倒映积水，撑透明伞的少女",
    "size": "2K",
    "n": 4
  }'
# → { "data": [ { "url": "https://…" }, … ] }

from openai import OpenAI
client = OpenAI(base_url="https://api.ofox.ai/v1", api_key="$OFOX_API_KEY")

res = client.images.generate(
    model="google/gemini-3-pro-image",
    prompt="赛博朋克东京雨夜，霓虹倒映积水，撑透明伞的少女",
    size="2K", n=4,
)
print(res.data[0].url)

import OpenAI from "openai";
const client = new OpenAI({ baseURL: "https://api.ofox.ai/v1", apiKey: process.env.OFOX_API_KEY });

const res = await client.images.generate({
  model: "google/gemini-3-pro-image",
  prompt: "赛博朋克东京雨夜，霓虹倒映积水，撑透明伞的少女",
  size: "2K", n: 4,
});
console.log(res.data[0].url);

The sample uses the OpenAI-compatible endpoint (swap the model to switch vendors); each provider also offers a native API, which usually fits the model's capabilities more closely. API docs: OpenAI Images API ↗ · Gemini native API ↗

FAQ

Frequently asked questions

Which models are supported?

Google Gemini 3, OpenAI GPT Image, Volcengine Doubao Seedream, Alibaba Qwen-Image, and Microsoft MAI Image — 14 models across 5 leading providers, with more added continuously. Switch between them through one API, compare the same prompt side by side, and pick the best fit.

Do you support image-to-image / reference images?

Yes. Passing a reference image routes to the image editing endpoint (image-to-image) for style continuation, inpainting, and multi-image fusion; natively supported by gpt-image, Gemini, and Seedream.

What output resolutions can I get?

Standard / 2K / 4K tiers covering aspect ratios like 16:9, 1:1, and 9:16, with batch output per request; each model's native maximum resolution in one step.

Is there a queue for generation?

Requests are processed asynchronously — long tasks return results via polling or callback without blocking, with stable scheduling under high concurrency.

Is my data secure?

All requests and generated data are transmitted over HTTPS. Results are never used for model training or public indexing — the content is yours.

5 providers · 14 image models

Start generating images today

Grab a key and call your first model in minutes — every top image model, one integration.

Get API key →