What the test measures
FluentCheck measures written comprehension through multiple-choice questions linked to practical and informational passages. The current release does not measure speaking, listening, pronunciation or spontaneous writing.
How questions are selected
Every attempt draws thirty unique questions, with ten items from each of three relative difficulty bands. The selection and answer order are shuffled so a new attempt is not a fixed copy of the previous one.
Scoring and indicators
Answers are scored on the server after submission. The overall score and the three indicators summarise performance across more accessible, intermediate and more demanding reading items in that attempt.
How to read the level
Score bands provide an estimated level from A1 to C2 for orientation. This mapping is practical guidance, not proof that the participant has officially achieved a CEFR level.
Limits and review
Topic familiarity, concentration, device and question selection may affect the outcome. Questions, translations, accessibility and technical behaviour are reviewed over time, while a recognised examination remains necessary for formal certification.
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.