I needed to classify a lot of spanish words into different cefr levels and thought why not use it to quickly estimate your cefr level.
Drawing inspiration from the concept of ELO and calibration games: you pick your starting level, then if you know a word you move up a level, and if you don't know a word you move down a level. With 10 calibration words you'll know more or less your success rate at various vocabulary levels.
Advantages:
* Fast
* Retestability - a large database of words can be created using chatgpt
Disadvantages:
* Only tests vocabulary which may not be correlated with your other language skills
There are many improvements I want to add (eg. clustering words to improve categorization, a personalized vocabulary database for language learners) but it's already taken me too long. I hope this isn't to basic to post.