xcam.aiIndependent benchmark lab
Pre-launch · 2026-10-28

Independent measurements

Independent measurements of AI companion models.

A live arena for companion and roleplay models. Blind preference votes, time to first response and cost per conversation — with the method and the data published.

In build. First results are not published yet.

What will be measured

method in review
Planned measures. No benchmark results exist yet; this table is not a leaderboard.
MeasureReported asStatus
Head-to-head preferenceblind votes planned
Time to first responseseconds, repeated runs planned
Cost per conversationUSD, per model version planned
Method and raw dataversioned pages planned

Referee only

No own model. No companion product. No product to sell. If affiliate links ever exist, they are disclosed at the link and never change a ranking.

Measured, not marketed

Every number will carry a unit, a date, a model version and an uncertainty. The method, the limits and the raw data are part of the result, not a footnote.