Topic starter
25/10/2026 11:11 am
TL;DR: A turn-0 model selector can distinguish hard from easy sessions moderately well, but it misses the long, expensive tail where many tasks drift from cheap prompts into multi-day, high-cost work, according to Archestra’s analysis of text-only routing for coding agents. The result is a routing strategy that looks efficient by session count yet loses money once cache effects, rework, and human cleanup are counted.
Editorial analysis by NHI Mgmt Group, based on content published by Archestra: “Routing coding agents is harder than it looks”.
Key questions
Q: What breaks when coding-agent routing is based only on the opening prompt?
A: Prompt-only routing breaks when the task expands after the first turn, because the classifier never sees the repository, hidden dependencies, or later rework.
Q: Why can a cheaper model create higher total cost in a coding workflow?
A: A cheaper model can become more expensive when it produces long responses that expand the conversation history and raise input costs on later turns.
Q: How should teams decide when to hand work to a subagent?
A: Use a subagent only when the work is a closed stretch whose output can stand on its own and will not need heavy parent-context reasoning later.
Practitioner guidance
- Map routing to task state, not just prompt text Require repository signals, change scope, and session history before a model tier is chosen for coding-agent work.
- Instrument session cost and context growth Track how quickly sessions expand after the first turn, and compare that to the model tier originally selected.
- Separate closed work from parent-context work Delegate only stretches that end in a self-contained result and do not need to be re-opened by the parent agent.
Bottom line: Prompt-only model routing can look efficient while still missing the sessions that become long, expensive, and operationally messy.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →