First Xiaomi MiMo API Call | Generated by AI

Home 2026.04

Question: What is the Xiaomi MiMo API Open Platform, and how do I make my first API call using it?

Answer:

Introduction to the Xiaomi MiMo API Open Platform

platform.xiaomimimo.com is Xiaomi’s official API console for accessing the MiMo family of large language models. It is designed to be OpenAI-compatible, meaning you can use the same patterns you already know from the OpenAI SDK or any OpenAI-compatible client — just swap the base URL and API key.

What Is MiMo?

MiMo is Xiaomi’s AI model series, developed under AI lead Fuli Luo (formerly of DeepSeek). The platform currently supports all 8 models in the MiMo-V2.5 and MiMo-V2 series, with token-based billing and transparent, controllable quotas.

Key models available:

Model Highlights
mimo-v2.5-pro 1.02T-parameter MoE with 42B active parameters, hybrid-attention architecture, 1M-token context window, strongest for agentic/coding tasks
mimo-v2.5 310B-parameter Sparse MoE (15B active), trained on 48T tokens, native multimodal (image, video, audio, text), 1M context
mimo-v2-pro Older flagship with 1M token context, strong reasoning
mimo-v2-flash Fast, cost-efficient; 262,144 token context window, 8192 max output tokens
mimo-v2-omni Multimodal (text + image), 262,144 token context
mimo-v2-tts / mimo-v2.5-tts Text-to-speech via the chat completions endpoint

Base URL & Authentication

The base URL for the API is https://api.xiaomimimo.com/v1 and it uses OpenAI-compatible completions (openai-completions).

Authentication is done via an API key in the Authorization header: Authorization: Bearer <your-api-key-from-platform>

You create your API key inside the Xiaomi MiMo console at platform.xiaomimimo.com.

Making Your First API Call

Option 1: cURL

curl --location --request POST 'https://api.xiaomimimo.com/v1/chat/completions' \
  --header "Authorization: Bearer $MIMO_API_KEY" \
  --header "Content-Type: application/json" \
  --data-raw '{
    "model": "mimo-v2-flash",
    "messages": [
      {
        "role": "system",
        "content": "You are MiMo, an AI assistant developed by Xiaomi."
      },
      {
        "role": "user",
        "content": "Hello, please introduce yourself."
      }
    ],
    "max_completion_tokens": 1024,
    "temperature": 0.8,
    "top_p": 0.95,
    "stream": false
  }'

Option 2: Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_MIMO_API_KEY",
    base_url="https://api.xiaomimimo.com/v1"
)

response = client.chat.completions.create(
    model="mimo-v2-flash",
    messages=[
        {"role": "system", "content": "You are MiMo, an AI assistant developed by Xiaomi."},
        {"role": "user", "content": "Hello, please introduce yourself."}
    ],
    max_tokens=1024,
    temperature=0.8,
    top_p=0.95
)

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

Option 3: JavaScript (fetch)

const response = await fetch('https://api.xiaomimimo.com/v1/chat/completions', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_MIMO_API_KEY',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    model: 'mimo-v2-flash',
    messages: [
      { role: 'system', content: 'You are MiMo, an AI assistant developed by Xiaomi.' },
      { role: 'user', content: 'Hello!' }
    ],
    max_tokens: 1024,
    temperature: 0.8
  })
});

const data = await response.json();
console.log(data.choices[0].message.content);

Key API Parameters

Parameter Description
model Model ID (e.g., mimo-v2-flash, mimo-v2.5-pro)
messages Array of {role, content} objects (system, user, assistant)
max_completion_tokens Max tokens to generate
temperature Randomness (0–1); 0.8 recommended
top_p Nucleus sampling; 0.95 recommended
stream true for streaming via server-sent events
thinking Enable/disable chain-of-thought reasoning (supported on pro/v2.5 models)

Thinking / Chain-of-Thought Mode

For models like mimo-v2.5-pro, mimo-v2.5, mimo-v2-pro, and mimo-v2-omni, chain-of-thought (thinking) is enabled by default. You can control this with "enabled" or "disabled" options.

When using enable_thinking: true with multi-turn tool calls, the model returns both reasoning_content and tool_calls fields. You must persist all reasoning_content from previous turns in the messages array for subsequent requests to maintain reasoning continuity.

Pricing / Token Plan

The Token Plan supports monthly/annual subscriptions and covers all 8 models in the MiMo-V2.5 and MiMo-V2 series, with four tiers available for individuals to enterprises. It is compatible with mainstream development toolchains like OpenCode, OpenClaw, and Claude Code.

References:

Back Donate