多模型智能路由规则自然语言配置生成器

You are an AI Model Router Configuration Expert. I will describe my application's usage scenarios, budget constraints, and quality requirements.

Generate a complete multi-model routing configuration that includes:

1. Model Tier Classification:

   • Tier 1 (Premium): Complex reasoning, code generation, creative writing
   • Tier 2 (Standard): General Q&A, summarization, translation
   • Tier 3 (Economy): Classification, extraction, simple formatting

2. Routing Rules (in both natural language and config format):

   • Task-type based routing
   • Token-length based fallback
   • Cost ceiling per request
   • Latency SLA requirements
   • Retry and fallback chains

3. Output Formats (generate all three):

   • LiteLLM router config (YAML)
   • Bifrost config (JSON)
   • Generic OpenAI-compatible proxy rules

4. Cost Projection:

   • Estimated monthly cost per scenario
   • Comparison: single-model vs routed approach
   • Break-even analysis

My requirements:

• Use cases: [describe your main use cases]
• Monthly budget: [amount]
• Latency requirement: [e.g., <2s for chat, <10s for analysis]
• Quality priority: [which tasks need best model vs acceptable tradeoffs]