Where this term shows up here## This glossary entry is tied to the channel's live model-testing material, specifically the video Live Vibe Check: OpenAI's Super-fast Spark Model Running at 1,000 TOKENS / S[^8a401be3@0]. In that context, the orchestrator agent is the role responsible for coordinating how a model is put to work during a structured evaluation: it decides what to send to the model, sequences those interactions, and brings results back together so a comparison stays consistent and comparable rather than ad hoc.## Why the role matters## When the Every team runs a vibe check — a live or week-long side-by-side test of a flagship model — someone or something has to keep the evaluation moving: feeding prompts, handling the model's outputs, and keeping the same method reusable across models. The orchestrator agent is the term for that coordinating component. Without it, each model test would be a one-off; with it, the same structured test can be applied again and again[^8a401be3@0].## Key points## - An orchestrator agent coordinates a model's interactions during a structured test or workflow.## - In this channel, the term is used in the context of live vibe checks, where the goal is a consistent, reusable comparison method[^8a401be3@0].## - The concept connects directly to the broader topic of AI Model Vibe Checks and Comparisons, where the emphasis is on methods you can run yourself.