model routing

Agents re-send growing context every step, so cost scales worse than linearly. The signal is knowing where the tokens go and the levers (caching, model routing, context trimming, step caps) that shrink the bill.

Agent latency is a step-count problem disguised as a speed problem. The winning move is removing round trips, not shaving milliseconds off each one, and knowing which step you can delete is the whole interview.