Kimi-K2-Thinking API
CHATCODELLM

Kimi-K2-Thinking API

All You Need To Know About This Model

Overview

Model Provider:Moonshot AI
Model Type:Chat
State:Coming soon

Key Specs

Quantization:bf16
Parameters:1T
Context:256K
Pricing:$0.48 input / $2.00 output
Try Model
Quick Start
Reserve Dedicated Endpoint

Introduction

Kimi-K2-Thinking is Moonshot AI's latest open-source thinking agent, engineered for exceptional long-horizon reasoning. It deeply integrates step-by-step thought with dynamic tool use, maintaining stable performance across 200-300 sequential calls to set new state-of-the-art records on benchmarks like HLE and BrowseComp.

Powered by a trillion-parameter MoE architecture with a 256k context window, its native INT4 quantization delivers a 2x inference speedup, effectively combining deep reasoning capabilities with high efficiency.

Kimi-K2-Thinking API Usage

Model

Endpoint

moonshotai/kimi-k2-thinking


        1
        curl -X POST https://inference.canopywave.io/v1 \
      
        2
        -H "Content-Type: application/json" \
      
        3
        -H "Authorization: Bearer $CANOPYWAVE_API_KEY" \
      
        4
        -d '{
      
        5
          "model": "moonshotai/kimi-k2-thinking",
      
        6
          "messages": [
      
        7
            {"role": "system", "content": "You are a helpful assistant."},
      
        8
            {"role": "user", "content": "please tell me a story."}
      
        9
          ],
      
        10
        }'
Contact us

Hi. Need any help?