Skip to content
TheChatArchiveThe Chat Archive
Geminigemini-2.5-flash· GoogleHugging Face LMArena Human Preference 140K

correct English The aim of this study is to identify the opportunities and limitations…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Claudeclaude-opus-4-20250514’s answer.
  1. User

    correct English The aim of this study is to identify the opportunities and limitations of using ChatGPT, Perplexity.ai, and GigaChat for creating English - language exercises. To strengthen the focus on results (rather than the research process itself), the goal emphasizes evaluating how effectively these generative AI tools enable the successful development of English exercises. The effectiveness of creating exercises with generative models was evaluated by experts based on three criteria: the match between the model - generated output and the initial request, the ease of creating exercises, and the need for subsequent editing by teachers. The scientific novelty of this research lies in assessing the effectiveness of ChatGPT, Perplexity.ai, and GigaChat for the specific task of English exercise creation—with expert - driven criteria measuring how well these tools perform in this educational context.

  2. Gemini Assistant

    The corrected text is as follows:

    The aim of this study is to identify the opportunities and limitations of using ChatGPT, Perplexity.ai, and GigaChat for creating English language exercises. To strengthen the focus on results (rather than the research process itself), the goal emphasizes evaluating how effectively these generative AI tools enable the successful development of English exercises.

    The effectiveness of creating exercises with generative models was evaluated by experts based on three criteria: how well the model-generated output matched the initial request, the ease of exercise creation, and the extent of subsequent editing required by teachers. The scientific novelty of this research lies in assessing the effectiveness of ChatGPT, Perplexity.ai, and GigaChat for this specific task—using expert-driven criteria to measure their performance in an educational context.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.