Prompt Routing
Directing an incoming request to the most appropriate AI model, agent, or workflow based on the request's characteristics—complexity, domain, required capabilities, cost sensitivity, and latency requirements. A prompt router might send simple FAQ questions to a small, fast model (Haiku-class) and complex reasoning tasks to a large, capable model (Opus-class), optimizing the cost-quality tradeoff across the full spectrum of requests. Advanced routing considers user tier, request urgency, and current system load.
Written by Max Zeshut
Founder at Agentmelt · Last updated Oct 26, 2026
Example
A customer support system receives 1,000 daily queries. The prompt router sends 600 simple questions (password resets, order status) to a small model at $0.001/query, 350 moderate questions to a mid-tier model at $0.01/query, and 50 complex escalations to a frontier model at $0.05/query—achieving 90% cost savings versus routing everything to the frontier model.
Frequently asked questions
- How does prompt routing differ from intent classification?
- Intent classification identifies what the user wants (refund, question, complaint). Prompt routing decides which model or agent should handle it. They often work together: intent classification determines the task type, and prompt routing selects the optimal model based on that type's complexity and requirements. Routing also considers operational factors like cost, latency, and current model availability.