From curious to confident in one sentence

The gap between curious and confident is usually configuration. You want to know how a few models handle your prompt, but answering that means picking models, choosing a scoring method, and setting it up before you see anything. The setup stands between you and the first result, so the first result often never happens.

A curated starter preset closes that gap. It ships ready to run, so your first comparison is a single sentence with nothing to configure.

The one-sentence first run

The curated preset is called “Quality spread”: a representative set of four low-cost models, saved and ready. You supply a prompt, it supplies everything else.

Ask your agent to run it on a prompt you actually care about, and the side-by-side comes back: how four models handle your words, and what each one costs. That is the whole aha moment, the shape of a result, delivered before you have built anything of your own.

Then make it yours

Seeing a result is what makes the next step obvious. Once you have watched the curated preset run, you know what you would change: your models, your scoring, your rubric. Build that run once and save it as your own preset, and your second question is as fast as your first.

That is the arc a curated preset is designed for. It takes you from “I wonder how these models do on my prompt” to a result in one sentence, and from there to your own saved standard. Curious to confident, without a configuration screen in the way.

Presets are on Pro and above. The curated starter preset is ready now; your own is one saved run away.