Instead of guessing which model will give the best answer, you can send one prompt to two or more models at once and see their responses side by side — right in the session, with your team.
From the composer at the bottom of any session, click Compare models instead of picking a single one. Choose which models to include, then send your message as usual.
The main risk is anchoring — a visible top tier can suppress mid-tier conversion...
I'd frame this around perceived value jump rather than the raw price delta...
Anyone in the session can mark a response as the preferred one — that selection is recorded, so if someone asks later “why did we go with this framing,” the comparison and the choice are both still there in the Timeline, not lost once the conversation moves on.
You don't have to compare every time. Pick a default model from the dropdown next to the composer, and switch it mid-session whenever you want — every switch is logged as a Timeline event, so it's clear which model produced which response later.
Was this article helpful?