Analyzing LLM Performance: A Comparative Study of ChatGPT and Gemini with Markdown History
Published:Jan 13, 2026 22:54
•1 min read
•Zenn ChatGPT
Analysis
This article highlights a practical approach to evaluating LLM performance by comparing outputs from ChatGPT and Gemini using a common Markdown-formatted prompt derived from user history. The focus on identifying core issues and generating web app ideas suggests a user-centric perspective, though the article's value hinges on the methodology's rigor and the depth of the comparative analysis.
Key Takeaways
- •The article proposes using Markdown to format chat histories for LLM comparison.
- •It aims to identify a user's key problems and compare the strengths of different LLMs (ChatGPT, Gemini).
- •It includes instructions, templates, and emphasizes the importance of masking personal/sensitive information.
Reference
“By converting history to Markdown and feeding the same prompt to multiple LLMs, you can see your own 'core issues' and the strengths of each model.”