Free tool · powered by the Calibration Judge

Ask once. Watch the AIs drift.

One question goes to several AI models at the same time. Then the Calibration Judge scores every answer — accuracy, bias, confidence — and shows exactly what's drifting.

Try Confidence Check with a friend →

Pick a challenge

Can you catch an AI bluffing?

Start with a question. See who knows their limits.

24 prompts
Browse all 24 questions

A good answer can be “I don’t know.” These challenges don’t guarantee a mistake.

Pick one, edit it if you like, then run the Arena.

0 / 500
The arena is idle. Ask something and hit Run the Arena — the models enter, one judge scores.

Educational & entertainment tool. The judge is also an AI and can be wrong — treat scores as a conversation starter, not a verdict. No financial advice. $CALL is a community token with a thesis; this tool exists because the thesis deserves to be tested in public.

Built by @Callibrecoin · callibrecoin.com · Community