AMD Radeon Cloud

AMD’s Token Factory: five shared free APIs and $10 of usage every day, reset at midnight.

AMD Radeon Cloud’s Token Factory is a shared OpenAI-compatible API on AMD GPUs. Log in at developer.amd.com.cn/radeon/tokenfactory, open a Public Free Model API card, and copy the key. Base URL is developer.amd.com.cn/radeon/api/v1. The five shared ids today are DeepSeek-V4-Flash-0731, DeepSeek-V4-Flash-Vision-Exp, Qwen3.8-Flash-Next, MiniCPM-V46 and MiniCPM5-1B.

The daily API quota is $10 USD, which at current rates is about 10 million to 111 million input/output tokens, depending on the model. It resets every day. Dedicated Model APIs on the same page are not that pool — they spend account compute. DeepSeek V4 Flash still has 1M of context. No card to start.

Free models: 5
Max context: 1M
Free tier: $10 of usage per day
Requirement: Email account, no card

Endpoint

Base URLhttps://developer.amd.com.cn/radeon/api/v1

Env varAMD_API_KEY

Rate limit$10 of usage per day (~10M–111M tokens, resets daily)

OpenAI-compatible
text
image
vision

Code examples

npm install openai

import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://developer.amd.com.cn/radeon/api/v1',
  apiKey: process.env.AMD_API_KEY
})

const response = await client.chat.completions.create({
  model: 'DeepSeek-V4-Flash-0731',
  messages: [{ role: 'user', content: 'Explain closures in one paragraph.' }]
})

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

pip install openai

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://developer.amd.com.cn/radeon/api/v1",
    api_key=os.environ["AMD_API_KEY"],
)

response = client.chat.completions.create(
    model="DeepSeek-V4-Flash-0731",
    messages=[{"role": "user", "content": "Explain closures in one paragraph."}],
)

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

curl https://developer.amd.com.cn/radeon/api/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AMD_API_KEY" \
  -d '{
    "model": "DeepSeek-V4-Flash-0731",
    "messages": [{"role": "user", "content": "Explain closures in one paragraph."}]
  }'

Worth knowing

The five Public Free Model APIs are the shared endpoints. Dedicated Model APIs spend account compute credits.
DeepSeek-V4-Flash-Vision-Exp is marked Limited Free. The standing free text ids are DeepSeek-V4-Flash-0731 and Qwen3.8-Flash-Next.
The portal lives on developer.amd.com.cn. Click a free card to copy the key and the curl sample.

Models you get for free5

DeepSeek V4 FlashDeepSeek-V4-Flash-0731
DeepSeek V4 Flash VisionDeepSeek-V4-Flash-Vision-Exp
Qwen 3.8 Flash NextQwen3.8-Flash-Next
MiniCPM V 4.6MiniCPM-V46
MiniCPM 5 1BMiniCPM5-1B