Kimi K3.1 Pro

Kimi K3.1 Pro is the first model in the K3.1 series newly released by Moonshot AI, and it is Kimi's most capable 3-trillion-parameter-class open-weight flagship model to date. It features 2.8 trillion (2.8T) MoE parameters, built upon the Kimi Delta Attention (KDA) hybrid linear attention architecture and Attention Residuals technology, natively supporting visual comprehension and providing an ultra-long context window of 1 million tokens (1,048,576 tokens).

Designed specifically for real-world frontier Agent tasks, this nearly 3-trillion-parameter model achieves a significant leap in long-horizon full-project programming, parametric 3D creation and modeling, end-to-end financial research and modeling, and complex long-horizon agentic workflows, while substantially reducing hallucination rates in complex reasoning.

Base URL

https://api.icreat.ai/v1

Model Code

kimi-k3.1-pro

Authentication

All API requests require authentication via an API Key. You can obtain your API Key from the console.

export ICREAT_API_KEY="your-api-key-here"

HTTP Request Headers

import os

API_KEY = os.environ.get("ICREAT_API_KEY")
headers = {
    "Content-Type": "application/json",
    "Authorization": "Bearer " + API_KEY,
}

Keep Your API Key Secure Never expose your API Key in client-side code or public repositories. Use environment variables or a backend proxy.

Code Examples

This model is accessed via an OpenAI-compatible Chat Completions endpoint and supports both streaming and non-streaming modes. Setting stream to false (default) in the request body causes the server to return the complete JSON response at once; setting it to true streams incremental chunks using Server-Sent Events (SSE), suitable for real-time output display.

POST/chat/completions

curl --fail-with-body --connect-timeout 10 --max-time 60 \
  -X POST https://api.icreat.ai/v1/chat/completions \
  -H "Authorization: Bearer ${ICREAT_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "kimi-k3.1-pro",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Coffee or tea ? Which is healthier? "}
  ],
  "thinking": {"type": "enabled"},
  "reasoning_effort": "low",
  "stream": false
}'

Input Parameters

The following parameters are accepted in the request body.

Total: 7 Required: 2 Optional: 5

modelstringrequired

The model ID used for completion. Must be the iCreat model_code (kimi-k3.1-pro for this model, not the original provider model name).

Example: "kimi-k3.1-pro"

messagesarray[object]required

The list of messages comprising the conversation.

max_tokensinteger

The maximum number of tokens to generate.

temperaturenumber

Sampling temperature, between 0 and 2.

streamboolean

If set to true, streams the response via Server-Sent Events (SSE).

thinkingobject

Extended thinking mode configuration (K3.1 Pro enables reasoning by default).

reasoning_effortstring

Reasoning intensity control. Supports low, high, and max (max is the default). For simpler tasks, you can lower the reasoning intensity to balance quality, latency, and token consumption.

lowhighmax

Output Parameters

The API returns a response format compatible with OpenAI Chat Completions.

Total: 6

idstring

A unique identifier for the completion.

objectstring

The object type, always chat.completion.

createdinteger

The Unix timestamp of when the completion was created.

modelstring

The model ID used for the completion.

choicesarray[object]

A list of completion choices.

usageobject

Token usage statistics.

LLM-Friendly Prompt

Below is an LLM-friendly Markdown prompt that can be copied into AI assistants such as Cursor or ChatGPT to help the AI understand this model's API integration method, invocation workflow, and key parameters. Click the "Copy LLM Prompt" button or copy the full text from the code block below.

# kimi-k3.1-pro

> Kimi K3.1 Pro is the latest 3-trillion-parameter-class (2.8T MoE) open-weight flagship model released by Moonshot AI, natively supporting visual comprehension and a 1-million-token context window, purpose-built for long-horizon programming, 3D creation, financial research, and complex Agent workflows.

## Overview

Use iCreat's OpenAI-compatible Chat Completions API for conversational completions, supporting streaming, non-streaming, extended thinking mode, and reasoning intensity configuration.

## API Information

- **Base URL**: `https://api.icreat.ai/v1`
- **Endpoint (POST)**: `/chat/completions`
- **Model ID**: `kimi-k3.1-pro`
- **Authentication**: `Authorization: Bearer ${ICREAT_API_KEY}`

## Invocation Workflow

Single POST request; `stream: false` returns a complete JSON response, and `stream: true` streams via SSE.

### Input Notes

- `model` (required): iCreat model_code `kimi-k3.1-pro`
- `messages` (required): List of conversation messages
- Common optional fields: `max_tokens` (maximum output tokens), `temperature` (sampling temperature), `stream` (streaming toggle), `thinking` (extended thinking mode), `reasoning_effort` (reasoning intensity, supports `low`, `high`, `max`)

### Output Notes

- Read the completion content from `choices[0].message.content`

## Important Notes

- `model` must strictly match the iCreat model_code
- All other fields follow standard OpenAI Chat Completions protocol specifications