Comparison
Cascade / fallback chainvsModel routing
Cascade / fallback chain
you try the cheap model first and escalate to the strong one only when the cheap answer fails a check.
Try the cheap model first, escalate to a stronger one when the attempt fails validation or confidence checks. The sequential cousin of routing; needs a machine-checkable success signal to work honestly.
Full entry →Model routing
you send the easy requests to the small cheap model and only the hard ones to the expensive one.
Sending each request to the cheapest model adequate for it: simple queries to a small model, hard ones to a frontier model, based on rules or a lightweight classifier. Routine 60-to-90 percent cost reductions when the traffic mix is genuinely varied.
Full entry →