How we decide what a level means.

AI IQ estimates AI tool-use mastery — judgment, verification, privacy, and productive use of available tools. It is not a clinical intelligence score, and not a test of whether someone can train models or join an AI engineering team.

Scoring

AI IQ ≈ 100 + (10 × level), with a small accuracy tweak. L1 Curious ≈ 110 … L7 Overlord Wrangler ≈ 170.

The seven levels

1. L1 Curious — Hasn't really started. No prerequisites, no shame.
2. L2 Hands On — Uses an assistant for real tasks, with mixed results.
3. L3 Capable — Gets useful output consistently on familiar work.
4. L4 Verifier — Checks output and catches errors before they travel.
5. L5 Builder — Turns a good result into a repeatable way of working.
6. L6 Coach — Can teach someone else and answer their questions.
7. L7 Overlord Wrangler — Sets direction for a team, and knows the limits.

Free short Check

Adaptive items get harder when answers are correct, until the system finds an edge. Unscored questions at the end capture which tools you have tried free, paid, or currently use.

Full Appraisal

Signed-in individuals may take a longer easy→hard battery free (self). Prior Check items are not repeated. Group members take the same instrument under an organizer's paid seats.

Fair challenge

A mid-range result after a fair test is useful. Getting stumped is measurement working — not a character judgment.

Privacy

Built for staff and leadership tool-use posture. Organizers can see results for people they bring into the system. See the privacy promise.