Scope
Questions are selected across three relative difficulty bands. Thirty unique items are drawn for each attempt, and scoring takes place on the server after submission.
How it works
Score bands provide an estimated proficiency level from A1 to C2. The mapping is guidance and does not mean the participant has officially achieved a CEFR level.
What to expect
Familiarity with a topic, concentration and test format can affect performance. An official examination remains necessary for immigration, university admission or formal certification.
Safeguards
Every attempt contains ten questions from each relative difficulty band. Questions are shuffled, answers are scored on the server and no answer key is exposed to the browser.
Your choices
The three indicators describe performance across easier, intermediate and more demanding items in this attempt. They are not separate psychometric subscales.
Contact
Question quality, translations and technical behaviour are reviewed over time. Reports about ambiguity or accessibility are investigated by the editorial team.
Question sources and limitations
The modern-language reading questions are adapted from Meta’s Belebele benchmark under CC BY-SA 4.0. The three difficulty bands are internal groupings, not validated CEFR levels.
The Latin test uses 60 original reading exercises drafted with AI assistance. They have not been independently validated by a Latin examiner; the result is an indicative self-assessment, not an accredited qualification.