A multi-agent research workflow currently runs four independent specialist agents one after another; total duration is the sum of all four. Model deployment rate limits still have headroom for higher concurrency. Product requires lower p95 latency without changing answer quality targets. What optimization should the team apply first?
Select an answer to reveal the explanation.
Short Explanation
If specialists are independent and you still have rate-limit headroom, run them at the same time, then merge the results. That cuts wall-clock time without changing the quality target. Extra sleeps, duplicate flooding, or a weaker tool-less model only make things worse.
Full Explanation
Increase task parallelism so independent specialists run concurrently, while respecting deployment rate limits and merging results afterward. Optimizing task duration for multi-agent solutions includes parallelism and rate-limit awareness. Independent specialists should run concurrently when dependencies allow, constrained by deployment quotas, with aggregation of outputs. Adding artificial sleep between every agent call increases latency. Flooding the endpoint with duplicate identical calls wastes tokens and risks throttling without improving useful work. Moving the entire workflow to a single smaller model that cannot use tools may harm quality and does not specifically exploit available parallelism. Exam focus: duration optimization means parallelizing independent work under rate limits.