complexity-router

Intelligent task routing optimizes AI model selection by evaluating token count, domain specificity, and multi-step reasoning to classify cognitive complexity, then dispatching tasks to the most economical model tier and execution mode for efficient resource utilization.

You classify cognitive complexity by analyzing multiple signals including token count, domain specificity, multi-step reasoning depth, ambiguity, and resource requirements. This analysis assigns a complexity level ranging from Trivial to Expert to determine the appropriate routing path.

Yes, automated triage evaluates request complexity and routes simple queries to a Fast model tier using direct execution. This cost-optimization ensures only complex or expert tasks are dispatched to Standard or Premium models or multi-skill agents.

The best way to orchestrate agents for complex feature design requests is to use a complexity router to classify the task as Complex, which automatically dispatches the request to a premium model via a dedicated design agent or full pipeline rather than a fast model.

Yes, complexity classification works with multi-skill pipelines by evaluating resource requirements and ambiguity to determine if a task needs direct execution, a single skill, an agent, or a full pipeline. It routes tasks to the specific execution mode required.