Artificial intelligence model routing selects among multiple available models for a particular request. A router can use task type, expected quality, response speed, price, data sensitivity or tool support to make the choice. The selection may be based on fixed rules, measured evaluations or another learned system.
Routing helps applications avoid using the most expensive model for every task. It also reduces dependence on one provider and allows teams to combine specialized models. Reliable routing requires current evaluation data, fallback behavior and monitoring because model performance and availability change over time.
Acronyms and aliases
AI model routing acronymmodel routing variant
General terms
Specialised terms
Related terms
Frequently asked questions
How does artificial intelligence model routing reduce cost?
It can send simple or low-risk requests to efficient models while reserving more capable and expensive models for tasks that actually need them.
What information can a model router consider?
A router may consider task category, quality targets, latency limits, price, context length, data policy, model availability and results from prior evaluations.
Videos explaining artificial intelligence model routing
