Technology
Students Prefer Google's Gemini Over Claude and ChatGPT for College Essays in Blind Tests
An analysis of nearly 7,000 blind student votes reveals that Gemini-family models are the top choice for writing and essay tasks, though performance varies by task stage.
The short version
- Google's Gemini emerged as the preferred AI family for college writing tasks, securing a 39.6% blind preference rate compared to Claude's 31.8% and OpenAI's 29.2%.
- The blind tests conducted by StudyArena showed that students tended to favor longer responses, but higher-effort reasoning settings actually decreased writing choice rates.
- While Gemini led for writing feedback, other AI families excelled at different stages, with Claude preferred for assignment planning and OpenAI models for research framing.
Key facts
- StudyArena analyzed 6,851 eligible, blind student votes in August 2026 to evaluate major AI provider families, including Gemini, Claude, and OpenAI's GPT family, for college writing and essay tasks.[Hacker News]
- For writing-specific tasks, Gemini achieved a 39.6% blind writing choice rate, ahead of Claude at 31.8% and ChatGPT/OpenAI models at 29.2%.[Hacker News]
- Selected answers were on average 37% longer than alternatives, and the longest response won 47.7% of decisive writing comparisons, while the shortest won 25.0%.[Hacker News]
- Higher-effort reasoning settings correlated with lower writing choice rates; 'low' settings earned a 40.7% choice rate compared to 29.5% for 'high' settings.[Hacker News]
- Provider families demonstrated distinct specialties, with Google's Gemini leading in writing feedback at 41.7%, Anthropic's Claude leading in assignment planning at 43.2%, and OpenAI's GPT family leading in research work at 39.3%.[Hacker News]
What remains uncertain
- The blind preference rankings are dynamic and may change over time as students cast new votes on StudyArena's live writing leaderboard.[Hacker News]