Why a model can score highly and still get things wrong.

High scores are averages over a test. Your question is one question.

Facts below come from the datasets on this site, last updated 29 Oct 2026, 14:36 UTC. Interpretation is labelled as ours. We have not tested these products hands-on.

What the evidence shows

• The highest SimpleQA Verified score among the listed models is 76.0%. That is the share it got right, so it missed the rest.
• Arena scores reflect votes. They do not check that an answer is correct.
• OTIS Mock AIME is close to its ceiling for leading models, so it separates them poorly: the best score is 100.0%.

Our interpretation

A score is an average over many questions. A model can be right most of the time and still wrong on the one you ask. Confidence in tone is not evidence of accuracy.

What to do next

Check anything important against a source you trust, and read how we test.