Qwen Image API guide

Use case

Developers searching for Qwen Image API access need one page that separates current public model aliases from provider-internal execution details.

Available models and pricing

Current Qwen Image access uses the public pricing rows for qwen-image-plus, qwen-image-max, and qwen-image-2.0.

• qwen-image-plus: $0.0286/image; public_endpoint_types = image-generation.
• qwen-image-max: $0.0714/image; public_endpoint_types = image-generation.
• qwen-image-2.0: $0.0286/image; public_endpoint_types = image-generation.

Customer-facing protocol

Crazyrouter exposes these Qwen Image models through the OpenAI Images-compatible POST /v1/images/generations route. Provider-side synchronous or asynchronous execution is handled behind the gateway.

• Use model qwen-image-max for a quality-first text-to-image baseline.
• Use size and n for standard OpenAI Images-compatible requests.
• Use response_format only where the Qwen docs allow url or b64_json.

When to compare Qwen Image

Qwen Image is useful when Chinese prompt handling, poster-like composition, or Alibaba-backed image generation is part of the evaluation set.

cn.crazyrouter.com test evidence

Only checks that returned 200 are shown here. API requests use https://cn.crazyrouter.com; account, billing, and console actions use https://crazyrouter.com.

Implementation examples

curl -X POST https://cn.crazyrouter.com/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "qwen-image-max",
    "prompt": "A premium product poster for a glass perfume bottle with soft studio lighting",
    "n": 1
  }'

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://cn.crazyrouter.com/v1",
)

response = client.images.generate(
    model="qwen-image-max",
    prompt="A premium product poster for a glass perfume bottle with soft studio lighting",
    n=1,
)

print(response.data[0].url)

FAQ