Skip to main content
OpenRouter is a unified API that provides access to multiple AI providers and models through a single interface. It’s perfect for users who want flexibility to try different models without managing multiple API keys and accounts.

​
Overview

  • Type: Cloud aggregator
  • Cost: Pay-per-use, varies by model (see pricing)
  • API Key Required: Yes
  • Installation Required: No
  • Official Website: https://openrouter.ai/

​
Prerequisites

1

Create an OpenRouter account

Sign up at openrouter.ai using your preferred authentication method.
2

Add credits (optional)

Navigate to Credits to add funds. OpenRouter offers free credits for trying models.
3

Generate an API key

Go to API Keys and create a new key. Copy it for use in AI Providers.
OpenRouter provides free credits when you sign up, allowing you to test different models before committing funds.

​
Setup in AI Providers

1

Select OpenRouter provider

In the AI Providers settings, click Create AI provider and select OpenRouter as the provider type.
2

Configure provider URL

Set the Provider URL to:
https://openrouter.ai/api/v1
3

Enter API key

Paste your API key from the API keys page into the API key field.
4

Select model

Click the refresh button to fetch the extensive list of available models, then select your preferred model (e.g., anthropic/claude-3.7-sonnet).
5

Test the provider

Click Test to verify your configuration works.

​
Available Models

OpenRouter provides access to models from multiple providers. Here are some popular options:

​
Premium Models

Model IDProviderDescription
anthropic/claude-3.7-sonnetAnthropicLatest Claude, excellent quality
openai/gpt-4oOpenAIGPT-4 Omni, multimodal
google/gemini-2.0-flash-expGoogleFast Gemini model
meta-llama/llama-3.3-70b-instructMetaPowerful open model
anthropic/claude-opus-4AnthropicHighest quality Claude

​
Cost-Effective Models

Model IDProviderDescription
openai/gpt-3.5-turboOpenAIFast and affordable
anthropic/claude-3-haikuAnthropicQuick responses
google/gemini-flash-1.5GoogleEfficient and fast
meta-llama/llama-3.2-3b-instruct:freeMetaFree tier option

​
Specialized Models

Model IDProviderDescription
perplexity/llama-3.1-sonar-huge-128k-onlinePerplexityWeb search integration
openai/o1-previewOpenAIAdvanced reasoning
qwen/qwen-2.5-72b-instructQwenExcellent multilingual
Model availability and pricing change frequently. Check the OpenRouter models page for the most current list.

​
Key Features

​
Single API for Multiple Providers

Access models from:
  • OpenAI
  • Anthropic
  • Google
  • Meta
  • Mistral
  • And many more
All through one API key and endpoint.

​
Free Models

Some models are available for free (with rate limits):
  • Look for models with :free in their ID
  • Perfect for testing and development
  • Community-supported options

​
Fallback Options

You can configure fallback models if your primary choice is unavailable or rate-limited.

​
Model Routing

OpenRouter can automatically route to the best available model based on:
  • Availability
  • Cost
  • Performance
  • Your preferences

​
Troubleshooting

​
Model Not Available

If a model shows as unavailable:
  1. Check the models page for current availability
  2. Some models require specific provider access
  3. Try an alternative model from the same family

​
Insufficient Credits

If requests fail due to credits:
  1. Check your credit balance
  2. Add more credits to your account
  3. Try a free model as an alternative

​
Rate Limits

Free models have rate limits. If you hit limits:
  • Wait for the limit to reset
  • Upgrade to a paid model
  • Use multiple free models in rotation

​
Model ID Format

Always use the full model ID format: 
provider/model-name
Examples:
  • anthropic/claude-3.7-sonnet
  • openai/gpt-4o
  • google/gemini-flash-1.5

​
Pricing Considerations

OpenRouter pricing varies by model:
  • Premium models: Similar to direct provider pricing (sometimes slightly higher)
  • Free models: No cost but rate-limited
  • Competitive models: Often better value than direct access
Cost-saving tips:

​
Advanced Configuration

​
Custom Headers

OpenRouter supports custom headers for:
  • App identification
  • User tracking
  • Request routing

​
Model Parameters

Pass standard parameters:
  • temperature - Control randomness
  • max_tokens - Response length
  • top_p - Nucleus sampling
  • stop - Stop sequences

​
Provider Preferences

You can specify provider preferences to:
  • Route to specific data centers
  • Require certain provider features
  • Optimize for cost or speed

​
Best Practices

  1. Start with free models: Test your integration without cost
  2. Compare models: Try different options for your use case
  3. Monitor costs: Check your usage regularly
  4. Set up fallbacks: Have backup models configured
  5. Use the right model: Balance cost, speed, and quality

​
Advantages of OpenRouter

  • Flexibility: Switch between models without changing code
  • Cost comparison: Easy to compare pricing across providers
  • No multiple API keys: Manage everything through OpenRouter
  • Community: Access to community-hosted models
  • Reliability: Automatic failover to alternative providers
OpenRouter is particularly useful if you want to experiment with different models or need fallback options when primary models are unavailable.

Build docs developers (and LLMs) love

Get started for freeTalk to us