OpenCode Proxy

Supported Models

MiMo v2.5 Thinking

Compact reasoning model optimized for math, code, and logical problem-solving. JanitorAI can show or hide thinking when available.

mimo-v2.5-free

DeepSeek V4 Flash Thinking

Fast inference model with strong coding and analytical performance. Thinking metadata is preserved for compatible clients.

deepseek-v4-flash-free

Nemotron 3 Ultra Thinking

High-performance model built for structured data extraction and complex reasoning with preserved thinking fields.

nemotron-3-ultra-free

North Mini Code North

Lightweight code-specialized model for quick completions, refactoring, and code explanation tasks.

north-mini-code-free

MiniMax M3 Nvidia

Large-scale reasoning model from MiniMax with strong multilingual and analytical capabilities.

minimax-m3

Step 3.7 Flash Nvidia

Fast inference model optimized for speed with solid reasoning and instruction-following performance.

step-3.7-flash

Kimi K2.6 Nvidia

Moonshot AI's flagship reasoning model with strong coding, math, and long-context capabilities.

kimi-k2.6

DeepSeek V4 Pro Nvidia

DeepSeek's most capable reasoning model for complex analysis, coding, and multi-step problem solving.

deepseek-v4-pro

Quick Start

# cURL
curl https://janitor-pi.vercel.app/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mimo-v2.5-free",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

# Python (openai SDK)
from openai import OpenAI

client = OpenAI(
    base_url="https://janitor-pi.vercel.app",
    api_key="anything"
)

response = client.chat.completions.create(
    model="deepseek-v4-flash-free",
    messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)

Showcase

Supported Models

Quick Start