Leaving exam simulator – Picture description (B2) GPT
Same flow as the standard simulator, but evaluation uses OpenAI Whisper for transcription and GPT-4o with the picture for feedback. In one run you work through 13 different pictures. Each round gives you 2 minutes to prepare and up to 75 seconds for a spoken description. After feedback you can move on to the next picture.
Speak fluently in English. Structure your description around: people, activities, place, time, atmosphere.
- Tenses: use Present Continuous for actions in progress and Present Simple for states and general facts.
- Position in the picture: use phrases like in the foreground, in the background, on the left / on the right.
- Speculation and opinion (important): use modal verbs (might, may, could, must, can't), seem / look, and opinion phrases (I guess, I believe, I suppose, I assume). Back up every speculation or opinion with what you can actually see (e.g. “I assume they are relaxed because they are smiling.”).
During preparation, mentally plan the order of your ideas; while recording, speak in connected sentences and do not jump between topics without linking words.