nirholas OpenRouter AI Models Guide
OpenRouter provides a unified API for 200+ AI models. This guide covers model selection, pricing, routing, and integration with SperaxOS.
What Is OpenRouter?
A single API endpoint that routes requests to 200+ models:
Your App → OpenRouter API → Model Provider
├── OpenAI (GPT-4, o1, o3)
├── Anthropic (Claude 4, Sonnet, Haiku)
├── Google (Gemini 2.5, Pro, Flash)
├── Meta (Llama 4, Maverick)
├── Mistral (Large, Medium)
├── DeepSeek (V3, R1)
└── 50+ more providers
Quick Start
curl https://openrouter.ai/api/v1/chat/completions \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "anthropic/claude-sonnet-4-20250514",
"messages": [{"role": "user", "content": "Hello!"}]
}'
Model Selection Guide
By Task
| Task | Recommended Model | Why |
|---|---|---|
| Complex reasoning | claude-sonnet-4, o3 | Best logical reasoning |
| Coding | claude-sonnet-4, deepseek-v3 | Best code generation |
| Creative writing | claude-sonnet-4, gpt-4o | Most creative output |
| Fast responses | gemini-flash, claude-haiku | Lowest latency |
| Long context | gemini-pro-2.5 (1M), claude-sonnet-4 (200K) | Largest context windows |
| DeFi analysis | claude-sonnet-4, gpt-4o | Best structured data handling |
| Vision/images | gpt-4o, gemini-pro | Best image understanding |
| JSON output | claude-sonnet-4, gpt-4o | Most reliable structured output |
| Budget | deepseek-v3, llama-4 | Cheapest per token |
By Price (per 1M tokens)
| Tier | Input | Output | Models |
|---|---|---|---|
| Free | $0 | $0 | Selected models with limits |
| Budget | $0.10-0.50 | $0.30-1.50 | DeepSeek, Llama, Gemma |
| Standard | $1-5 | $3-15 | GPT-4o, Claude Sonnet, Gemini Pro |
| Premium | $10-15 | $30-60 | o3, Claude Opus |
By Context Window
| Model | Context | Best For |
|---|---|---|
| Gemini 2.5 Pro | 1,048,576 | Entire codebases |
| Claude Sonnet 4 | 200,000 | Long documents |
| GPT-4o | 128,000 | Standard tasks |
| DeepSeek V3 | 64,000 | Budget long-context |
Routing Strategies
Automatic Routing
Let OpenRouter pick the best model:
{
"model": "openrouter/auto",
"messages": [...]
}
Fallback Chain
Define fallback models if primary is unavailable:
{
"model": "anthropic/claude-sonnet-4-20250514",
"route": "fallback",
"models": [
"anthropic/claude-sonnet-4-20250514",
"openai/gpt-4o",
"google/gemini-2.5-pro"
]
}
Cheapest Route
Use the cheapest available model for a task:
{
"model": "openrouter/cheapest",
"max_price": { "prompt": 0.001, "completion": 0.003 }
}
SperaxOS Integration
SperaxOS uses OpenRouter as a model provider:
Configuration
# .env
OPENROUTER_API_KEY=sk-or-...
OPENROUTER_MODEL_LIST=anthropic/claude-sonnet-4-20250514,openai/gpt-4o,google/gemini-2.5-pro
Model Data File
SperaxOS maintains a curated model list at data/openrouter-models.ts:
- Pre-selected models optimized for DeFi/crypto tasks
- Pricing and capability metadata
- Context window sizes
- Feature flags (vision, function calling, JSON mode)
Per-Agent Model Selection
Different SperaxOS agents can use different models:
- DeFi Advisor: Claude Sonnet 4 (best reasoning)
- Quick Chat: Gemini Flash (fastest response)
- Code Assistant: DeepSeek V3 (budget coding)
- Research Agent: Gemini 2.5 Pro (largest context)
API Compatibility
OpenRouter is compatible with:
- OpenAI SDK (just change base URL)
- LangChain
- LlamaIndex
- Vercel AI SDK
- Any OpenAI-compatible client
import OpenAI from 'openai';
const client = new OpenAI({
baseURL: 'https://openrouter.ai/api/v1',
apiKey: process.env.OPENROUTER_API_KEY,
});
Links
- Website: https://openrouter.ai
- Docs: https://openrouter.ai/docs
- Models: https://openrouter.ai/models
- GitHub: https://github.com/nirholas/openrouter-models
- SperaxOS: https://app.sperax.io