GPT-4o provided a structured breakdown with separate sections for REST and GraphQL. Covered key characteristics, HTTP methods, when to use each, and a bookstore CRUD example. Technically accurate but reads like documentation — jumps straight into definitions without building intuition. Doesn't demonstrate over-fetching/under-fetching problems.
Gemini opened with a restaurant analogy (REST = set menu, GraphQL = buffet with personal chef). Explicitly framed over-fetching and under-fetching problems. The practical example walked through the same data requirement (blog post + author + comments) solved both ways — showing 3 REST calls with extra unused data vs 1 GraphQL query with a perfectly shaped response. Ended with a comparison table for quick reference.
🔍 Analysis
These are real, unedited outputs from GPT-4o and Gemini 2.5 Pro given the exact same prompt.
Gemini's response is significantly better for a junior developer — the target audience specified in the prompt. It opens with a restaurant analogy (set menu vs buffet with a personal chef) that makes the abstract concept click instantly. It then explicitly frames the two core problems REST creates (over-fetching and under-fetching) before showing how GraphQL solves them. The practical example is exceptional: it walks through the same data requirement (blog post + author + comments) solved both ways, showing 3 REST calls with all the extra unused data vs 1 GraphQL query with a perfectly shaped response. The comparison table at the end provides a clean reference.
GPT-4o's response is technically accurate but reads more like documentation than teaching. It covers the same concepts but in a less memorable, less structured way.
Gemini wins on pedagogical quality — the analogy reduces cognitive load, the side-by-side practical example makes trade-offs visceral, and the comparison table provides an actionable reference.