Students prefer Gemini over ChatGPT and Claude for AI essays in blind tests

For college students, Gemini AI model has surpassed ChatGPT and Claude in producing preferred AI-assisted essays, according to a StudyArena study. In August 2026, StudyArena analyzed 6,851 blind student votes to determine which AI model was most favored for writing tasks. Gemini emerged as the top choice at 39.6%, followed by Claude at 31.8% and ChatGPT at 29.2%.

The study's blind test structure involved presenting the essay responses to students before revealing the AI model names, ensuring that external factors such as brand loyalty or prior experience did not influence their choices. Gemini's lead is significant because students had to select the best response solely based on the content displayed on the page.

The study suggests that Gemini's balance of completeness, structure, and voice makes it the ideal choice for college writing. College writing often calls for a balanced approach, rather than an extreme one, as the best response should be both complete and structured without sounding mechanical, and clear without losing the writer's voice.

Gemini's ability to meet these requirements made it the preferred model among students. The study also found that Gemini's answers were 37% longer on average than those from the other models. Longer responses tended to win more decisive writing comparisons, with the longest response winning 47.7% of the time and the shortest winning only 25.0%.

While Gemini's longer responses may seem excessive, the study notes that the extra length often allows for more points, qualifications, and repetition, which can be beneficial for research plans or proofreading. However, it is essential to avoid overdoing it, as excessive repetition can detract from the overall prose. For essay writing, the study recommends starting with a normal or low-effort setting on Gemini and only moving up the effort level when the challenging aspect of the essay is reasoning through evidence, not simply the writing process.

The final edit should focus on reducing the number of ideas and strengthening the best one. The current model families within StudyArena's writing data offer different reasoning and effort settings. Higher-effort labels did not necessarily result in higher choice rates; instead, more reasoning created room for more points, qualifications, and repetition, which can be helpful for research or proofreading but often hurt the flow of the essay.

For overall essay writing, Gemini remains the best first choice, as it can help identify generic passages, missing details, and weak structure. Once these issues are addressed, the user can then work on the language itself. While Claude remains useful for planning and generating alternative structures, Gemini's ability to improve finished pieces makes it the better overall choice for college students.