How to Access Free DeepSeek AI Models on China’s Supercomputing Center

Background

Recently I have been following DeepSeek updates, and now there is good news: the National Supercomputing Center has launched DeepSeek and provides free access to three models—DeepSeek‑R1:7B, DeepSeek‑R1:14B, and DeepSeek‑R1:32B. Below are the steps to use them.

Steps

Enter the navigation bar as shown and purchase the service.

After purchase, click “Use” to obtain the API key.

Test Code

The following Python script demonstrates how to call the DeepSeek‑R1‑Distill‑Qwen‑32B model via the provided endpoint.

import requests
import json

url = "http://activity.scnet.cn:61080/v1/chat/completions"
payload = json.dumps({
  "messages": [
    {
      "role": "user",
      "content": "写一首诗"
    }
  ],
  "stream": True,
  "model": "DeepSeek-R1-Distill-Qwen-32B",
  "temperature": 0.5,
  "presence_penalty": 0,
  "frequency_penalty": 0,
  "top_p": 1
}, ensure_ascii=False).encode('utf-8')
headers = {
  'Accept': 'application/json, text/event-stream',
  'Accept-Language': 'zh-CN,zh;q=0.9',
  'Authorization': 'Bearer',  # Insert access key
  'Connection': 'keep-alive',
  'Content-Type': 'application/json',
  'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.6261.95 Safari/537.36'
}
response = requests.request("POST", url, headers=headers, data=payload, stream=True)

for line in response.iter_lines():
    if line:
        print(line.decode('utf-8'))

Result

Running the script returns streamed responses from the model. The screenshot below shows the output.

Scan the QR code to join the technical discussion group.