Jev takes an IQ-style test
24 questions from warm-ups to multistep challenges. Watch Jev choose between plausible answers, then check every choice against the answer key.
Not a standardized IQ score. Estimate uses uncalibrated assumptions.
Run details and scoring limits
The local demo always picks A. It shows how scoring works and makes no model calls.
The numerical estimate uses assumed reference values; these original practice puzzles have no human population norms. Difficulty labels are editorial, not measured. Patterns are sent as text, not images.
Uniform random guessing would average 6 / 24 across repeated full tests. This is an expected value, not a measured run.
Answer sheet
Select a questionGrowing gaps
1 / 24What comes next? 6, 11, 18, 27, ?
Assumption Each successive gap is 2 larger than the previous gap.
Confidence describes the model’s response. The checked answer determines correctness.
Checked answer and explanation
B · 38
The gaps are 5, 7, 9, then 11. Adding 11 to 27 gives 38.