Qwen Vision

Qwen vision requests use the OpenAI-compatible Chat Completions API. Put text, images, and videos in the same messages[].content array, and call Cubicspaces /v1/chat/completions.

Endpoint

POST /v1/chat/completions

Supported Models

Model	Input	Output	Best for
qwen3.6-plus	Text, images, video	Text	Image and video understanding, text extraction, complex visual reasoning

Actual model availability depends on account permissions and platform configuration. qwen3.6-plus supports long-context multimodal requests and is suitable for multi-image, multi-page document, long-video, and visual reasoning workflows.

Image And Video Understanding

Single Image

curl https://cubicspaces.cloud/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "image_url",
            "image_url": {
              "url": "https://example.com/product.jpg"
            }
          },
          {
            "type": "text",
            "text": "Describe the main subject, scene, and likely use of this image."
          }
        ]
      }
    ]
  }'

Multi-Image Comparison

curl https://cubicspaces.cloud/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [
      {
        "role": "user",
        "content": [
          { "type": "image_url", "image_url": { "url": "https://example.com/before.jpg" } },
          { "type": "image_url", "image_url": { "url": "https://example.com/after.jpg" } },
          { "type": "text", "text": "Compare the two images and list the main changes." }
        ]
      }
    ]
  }'

Video Understanding

You can pass a publicly reachable video URL or a list of extracted frame image URLs. For long videos, streaming is recommended so the client does not wait on a single long response.

curl https://cubicspaces.cloud/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "video",
            "video": "https://example.com/demo.mp4"
          },
          {
            "type": "text",
            "text": "Summarize the key events in chronological order. Return a JSON array with start_time, end_time, and event."
          }
        ]
      }
    ],
    "stream": true,
    "stream_options": {
      "include_usage": true
    }
  }'

If your client has already extracted frames, pass a frame list:

{
  "type": "video",
  "video": [
    "https://example.com/frame-001.jpg",
    "https://example.com/frame-002.jpg",
    "https://example.com/frame-003.jpg"
  ],
  "fps": 2
}

Text Extraction

Use qwen3.6-plus to extract text or structured fields from images, screenshots, receipts, and tables. For machine-readable output, explicitly request JSON and define how missing fields should be represented.

curl https://cubicspaces.cloud/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "image_url",
            "image_url": {
              "url": "https://example.com/invoice.jpg"
            }
          },
          {
            "type": "text",
            "text": "Extract invoice_code, invoice_number, date, amount, and buyer_name from the image. Return JSON. Use null for fields that cannot be recognized."
          }
        ]
      }
    ],
    "enable_thinking": false,
    "response_format": {
      "type": "json_object"
    }
  }'

Visual Reasoning

qwen3.6-plus supports visual reasoning. For complex questions, chart analysis, object relationship reasoning, or video event attribution, use streaming and control reasoning output with enable_thinking.

curl https://cubicspaces.cloud/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "image_url",
            "image_url": {
              "url": "https://example.com/chart.png"
            }
          },
          {
            "type": "text",
            "text": "Analyze the trend in this chart, explain the likely cause, and give a conclusion."
          }
        ]
      }
    ],
    "enable_thinking": true,
    "stream": true
  }'

In streaming responses, read normal answer text from choices[0].delta.content. If the model returns reasoning, read it from choices[0].delta.reasoning_content.

SDK Examples

Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://cubicspaces.cloud/v1"
)

response = client.chat.completions.create(
    model="qwen3.6-plus",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://example.com/screenshot.png"
                    },
                },
                {
                    "type": "text",
                    "text": "Recognize the text in this screenshot and summarize it as bullet points."
                },
            ],
        }
    ],
    extra_body={
        "enable_thinking": False
    },
)

print(response.choices[0].message.content)

Node.js

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://cubicspaces.cloud/v1"
});

const response = await client.chat.completions.create({
  model: "qwen3.6-plus",
  messages: [
    {
      role: "user",
      content: [
        {
          type: "image_url",
          image_url: {
            url: "https://example.com/screenshot.png"
          }
        },
        {
          type: "text",
          text: "Recognize the text in this screenshot and summarize it as bullet points."
        }
      ]
    }
  ],
  enable_thinking: false
});

console.log(response.choices[0].message.content);

Notes

• Image and video URLs must be directly reachable by the platform service. Do not use URLs that require login, redirect to HTML pages, or block server-side downloads.
• Larger images and longer videos consume more input context. In production, compress images, trim videos, or extract key frames based on the task.
• enable_thinking is not a standard OpenAI parameter. With the OpenAI Python SDK, pass it through extra_body; with direct HTTP requests, place it at the top level of the JSON body.
• For JSON output, prefer disabling thinking mode and state field names, types, and missing-value rules in the prompt.