Auto-Routing

Automatic model selection based on task complexity.

How It Works

When you set model: "auto", the system analyzes your prompt and selects the optimal model based on:

1. Task Type - Code, analysis, creative, chat
2. Complexity - Simple vs complex reasoning
3. Cost Efficiency - Best value for the task
4. Response Quality - Match quality to requirements

Usage

{
  "query": "What is 2 + 2?",
  "model": "auto"
}

Response

The response includes routing information:

{
  "success": true,
  "response": "4",
  "model": "gpt-4o-mini",
  "provider": "openai",
  "routing": {
    "is_auto_routed": true,
    "model_chosen": "openai.gpt-4o-mini",
    "confidence": 0.92,
    "reason": "simple_query"
  }
}

Routing Logic

Task Type	Simple	Complex
Code	gpt-4.1-mini	grok-4-1-fast-reasoning
Analysis	claude-3-haiku	claude-opus-4-5
Creative	gpt-4.1	claude-sonnet-4-5
Chat	gpt-4o-mini	gpt-4.1

Task Recommendations

Get model recommendations for a specific task:

curl https://llm.zihin.ai/api/llm/recommendations/code_generation

Response:

{
  "success": true,
  "task": "code_generation",
  "recommendations": [
    {
      "id": "grok.grok-4-1-fast-reasoning",
      "name": "Grok 4.1 Fast Reasoning",
      "provider": "grok",
      "tier": "flagship"
    },
    {
      "id": "openai.gpt-4.1",
      "name": "GPT-4.1",
      "provider": "openai",
      "tier": "standard"
    }
  ]
}

Task Types

Task	Description
code_generation	Writing code
summarization	Condensing text
translation	Language translation
analysis	Data/text analysis
creative	Creative writing
chat	Conversational

Confidence Scores

The routing confidence indicates how certain the system is:

Score	Meaning
0.9+	High confidence
0.7-0.9	Good match
0.5-0.7	Acceptable
< 0.5	Consider specifying model

Override Auto-Routing

To bypass auto-routing, specify a model directly:

{
  "query": "Complex task here",
  "model": "claude-opus-4-5-20250514",
  "provider": "anthropic"
}