MiniMax models are fully OpenAI-compatible on Infercom. Use base URL api.infercom.ai/v1 with your existing OpenAI SDK - no code changes needed. MiniMax-M2.7 Ultraspeed and M2.5 are available with 400+ tokens/second inference speeds, EU-hosted in Munich for full GDPR compliance.
Available Models
All MiniMax models available on Infercom. Use the model ID in your API requests.
| Model | Model ID | Parameters | Context | Speed | Best For | |
|---|---|---|---|---|---|---|
| MiniMax-M2.7 Ultraspeed | MiniMax-M2.7 | 230B MoE | 192K tokens | 400+ tok/s | Self-critique, multi-agent, 30% better coding | View details → |
| MiniMax-M2.5 | MiniMax-M2.5 | 230B MoE | 160K tokens | 400+ tok/s | Agentic coding, cost-effective | View details → |
Model IDs are case-sensitive. Use exactly as shown.
Quick Start
Works with any OpenAI-compatible SDK. Just point to our endpoint.
Integration takes three steps: First, sign up at cloud.infercom.ai to get your API key. Second, change your base URL from api.openai.com to api.infercom.ai/v1. Third, set your model to MiniMax-M2.7 or MiniMax-M2.5. Your existing code, prompts, and tooling all work unchanged.
from openai import OpenAI
client = OpenAI(
base_url="https://api.infercom.ai/v1",
api_key="your-api-key"
)
response = client.chat.completions.create(
model="MiniMax-M2.7",
messages=[{"role": "user", "content": "Hello"}],
stream=True
)
for chunk in response:
print(chunk.choices[0].delta.content, end="")Supported Endpoints
POST /v1/chat/completionsOpenAI Chat Completions (streaming supported)
POST /v1/messagesAnthropic Messages API
POST /v1/responsesOpenAI Responses API
GET /v1/modelsList available models
Why Use Infercom for MiniMax
EU-Hosted
Munich datacenter. Full GDPR compliance. No US CLOUD Act exposure.
Up to 10x Faster
SambaNova's dataflow architecture delivers world-record inference speeds.
Zero Data Retention
Your prompts and outputs are never stored or used for training.
Drop-in Compatible
Works with OpenAI SDK, Anthropic SDK, LangChain, LlamaIndex, and any OpenAI-compatible tool.
Frequently Asked Questions
Is MiniMax API compatible with OpenAI?
Yes. MiniMax models on Infercom are fully OpenAI-compatible. Use base URL api.infercom.ai/v1 with your existing OpenAI SDK, LangChain, LlamaIndex, or any OpenAI-compatible tool. No code changes needed beyond the base URL.
What is the base URL for MiniMax API?
The base URL is https://api.infercom.ai/v1. Set this as your base_url in the OpenAI SDK or baseURL in JavaScript. Supported endpoints: /chat/completions (OpenAI), /messages (Anthropic), /responses (OpenAI Responses API), and /models.
Can I use LangChain with MiniMax?
Yes. Use ChatOpenAI from langchain_openai with base_url='https://api.infercom.ai/v1' and model='MiniMax-M2.7' or 'MiniMax-M2.5'. LlamaIndex, CrewAI, and other frameworks that support OpenAI also work.
What is the difference between MiniMax M2.7 and M2.5?
Both are 230B MoE models (10B active parameters). M2.7 Ultraspeed is the newer version with built-in self-critique, native multi-agent support, 192K context, and 30% better coding performance. M2.5 has 160K context and offers excellent agentic coding at lower cost. Both run at 400+ tokens/second on Infercom.
Is MiniMax API GDPR compliant?
Yes. MiniMax models on Infercom run on EU-hosted infrastructure in Munich, Germany. Full GDPR compliance with zero data retention - prompts and outputs are never stored or used for training. No US CLOUD Act exposure.