DashScope API Service

Overview

For production use without managing infrastructure, you can access Qwen3-VL through the DashScope API service. The API is fully compatible with the OpenAI client library, making integration seamless.

Getting Started

1. Get Your API Key

Obtain your DashScope API key from Aliyun Model Studio.

2. Install OpenAI Client

pip install openai

Using the API

Here’s a complete example of using the DashScope API with the OpenAI client:

from openai import OpenAI

# Set your DASHSCOPE_API_KEY here
DASHSCOPE_API_KEY = "your-api-key-here"

client = OpenAI(
    api_key=DASHSCOPE_API_KEY,
    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
)

completion = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=[{"role": "user", "content": [
        {"type": "image_url",
         "image_url": {"url": "https://dashscope.oss-cn-beijing.aliyuncs.com/images/dog_and_girl.jpeg"}},
        {"type": "text", "text": "这是什么"},
    ]}]
)
print(completion.model_dump_json())

API Configuration

Client Setup

When initializing the OpenAI client for DashScope:

client = OpenAI(
    api_key=DASHSCOPE_API_KEY,  # Your DashScope API key
    base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",  # DashScope endpoint
)

Available Models

The following Qwen3-VL models are available through the API:

qwen3-vl-235b-a22b-instruct: Large-scale instruction-tuned model
Other model variants (check the official documentation for the full list)

Multimodal Input

Image Input

Provide images via URL:

messages = [{
    "role": "user",
    "content": [
        {
            "type": "image_url",
            "image_url": {"url": "https://example.com/image.jpg"}
        },
        {
            "type": "text",
            "text": "Describe this image."
        }
    ]
}]

completion = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=messages
)

Video Input

Provide videos via URL:

messages = [{
    "role": "user",
    "content": [
        {
            "type": "video_url",
            "video_url": {"url": "https://example.com/video.mp4"}
        },
        {
            "type": "text",
            "text": "What happens in this video?"
        }
    ]
}]

completion = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=messages
)

Multiple Images

Include multiple images in a single request:

messages = [{
    "role": "user",
    "content": [
        {"type": "image_url", "image_url": {"url": "https://example.com/image1.jpg"}},
        {"type": "image_url", "image_url": {"url": "https://example.com/image2.jpg"}},
        {"type": "text", "text": "Compare these two images."}
    ]
}]

Response Format

The API returns responses in the standard OpenAI format:

completion = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=messages
)

# Access the response
response_text = completion.choices[0].message.content
print(response_text)

# Get full response as JSON
print(completion.model_dump_json())

Advanced Parameters

Customize generation with additional parameters:

completion = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=messages,
    max_tokens=2048,          # Maximum tokens to generate
    temperature=0.7,          # Sampling temperature (0.0-1.0)
    top_p=0.8,               # Nucleus sampling parameter
    stream=False             # Enable streaming responses
)

Streaming Responses

Enable streaming for real-time output:

stream = client.chat.completions.create(
    model="qwen3-vl-235b-a22b-instruct",
    messages=messages,
    stream=True
)

for chunk in stream:
    if chunk.choices[0].delta.content is not None:
        print(chunk.choices[0].delta.content, end="")

Error Handling

Implement proper error handling:

from openai import OpenAI, OpenAIError

try:
    completion = client.chat.completions.create(
        model="qwen3-vl-235b-a22b-instruct",
        messages=messages
    )
    print(completion.choices[0].message.content)
except OpenAIError as e:
    print(f"API error: {e}")

Best Practices

API Key Security

Never hardcode API keys in your source code

Use environment variables:

import os
DASHSCOPE_API_KEY = os.getenv("DASHSCOPE_API_KEY")

Rate Limiting

Implement exponential backoff for retries
Monitor your API usage
Cache responses when possible

Cost Optimization

Set appropriate max_tokens limits
Use batch processing for multiple requests
Monitor token usage in responses

Additional Resources

For more detailed information, refer to:

Next Steps

For self-hosted deployment, see vLLM deployment
Explore SGLang deployment as an alternative
Try Docker deployment for local testing

Get Started

Core Concepts

Inference

Deployment

Fine-tuning

Capabilities

DashScope API Service

Overview

Getting Started

1. Get Your API Key

2. Install OpenAI Client

Using the API

API Configuration

Client Setup

Available Models

Multimodal Input

Image Input

Video Input

Multiple Images

Response Format

Advanced Parameters

Streaming Responses

Error Handling

Best Practices

API Key Security

Rate Limiting

Cost Optimization

Additional Resources

Next Steps

​Overview

​Getting Started

​1. Get Your API Key

​2. Install OpenAI Client

​Using the API

​API Configuration

​Client Setup

​Available Models

​Multimodal Input

​Image Input

​Video Input

​Multiple Images

​Response Format

​Advanced Parameters

​Streaming Responses

​Error Handling

​Best Practices

​API Key Security

​Rate Limiting

​Cost Optimization

​Additional Resources

​Next Steps

Overview

Getting Started

1. Get Your API Key

2. Install OpenAI Client

Using the API

API Configuration

Client Setup

Available Models

Multimodal Input

Image Input

Video Input

Multiple Images

Response Format

Advanced Parameters

Streaming Responses

Error Handling

Best Practices

API Key Security

Rate Limiting

Cost Optimization

Additional Resources

Next Steps