Digitorn
Digitorn
All terms
Cost & ops

Model routing

Sending different turn types to different models (cheap for exploration, premium for writing).

also known as: per-agent brain, brain routing
In depth

Model routing is the single biggest cost lever in multi-agent setups. Half of an average coding session is exploration (grep, glob, read), which a small model handles fine. The writing and editing turns need a premium model. Routing each agent to the right model in YAML is two lines of config, the savings are typically 50-65% on real workloads.

Related concepts
Read the deep dive

How we cut our coding agent's bill by 60% with model routing

Read article
Newsletter

Get the next post in your inbox.

Engineering notes from the Digitorn team. No marketing, no launch announcements, no "10 prompts that will change your life". Just the things we write that we'd want to read.

One-click unsubscribe. We never share your address. Powered by our own infrastructure, not a tracker.

More in Cost & ops

Audit log/glossary/audit-logBrain fallback/glossary/brain-fallbackCost cap/glossary/cost-capPII redaction/glossary/pii-redactionRate limiting/glossary/rate-limitingTurn cap/glossary/turn-cap