Testing OpenRouter API: Free AI Model Access

Learn how to test and use free AI models through OpenRouter's unified API.

Prerequisites

• OpenRouter account (free at https://openrouter.ai)
• Python 3.8+ with requests: pip install requests
• API key from OpenRouter dashboard

Quick Test with curl

curl -X POST https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.0-flash-exp:free",
    "messages": [{"role": "user", "content": "Say hello in 3 languages"}]
  }'

Python Integration

import requests
import os

OPENROUTER_URL = "https://openrouter.ai/api/v1/chat/completions"

def test_openrouter(prompt: str, model: str = "google/gemini-2.0-flash-exp:free") -> str:
    """Test OpenRouter API with a prompt."""
    headers = {
        "Authorization": f"Bearer {os.getenv('OPENROUTER_API_KEY')}",
        "Content-Type": "application/json"
    }
    data = {
        "model": model,
        "messages": [{"role": "user", "content": prompt}],
        "temperature": 0.7,
        "max_tokens": 150
    }

    response = requests.post(OPENROUTER_URL, headers=headers, json=data, timeout=30)

    if response.status_code == 200:
        return response.json()["choices"][0]["message"]["content"]
    else:
        raise Exception(f"Error: {response.status_code} - {response.text}")

# Test it
if __name__ == "__main__":
    prompt = "Explain the difference between REST and GraphQL in 2 sentences"
    try:
        result = test_openrouter(prompt)
        print(f"Q: {prompt}")
        print(f"A: {result}")
    except Exception as e:
        print(f"Error: {e}")

Free Models Available

OpenRouter offers many free models (look for :free suffix):

Model	Best For
google/gemini-2.0-flash-exp:free	General chat, coding
meta-llama/llama-3.1-8b-instruct:free	Reasoning, analysis
mistralai/mistral-7b-instruct:free	Fast responses
microsoft/phi-3-mini-128k-instruct:free	Long context

Best Practices

1. Use environment variables for API keys
2. Monitor rate limits - Free tier has limits
3. Set max_tokens to control response length
4. Handle errors gracefully with retries
5. Test multiple models to find the best fit

Conclusion

OpenRouter provides a unified interface to dozens of free AI models. Perfect for testing, learning, and building AI-powered applications without API costs.