ChatGPT Thinking mode reshapes citations, visibility drops across modes
ChatGPT’s Thinking mode pulled citations from a different source set, with only 25.6% domain overlap and finance, health and SaaS gaining while Reddit lost share.
ChatGPT’s Thinking mode pulled from a different citation pool than its faster answer path, with only 25.6% overlap in cited domains across the same prompts. Semrush and Kevin Indig’s test found citation rates rose from 50% to 68% and sources per response climbed from 2.6 to 4.5 as reasoning increased.
The test ran 100 prompts twice, for 200 total responses, across 20 buyer journeys in four verticals. High-reasoning mode ran 1,130 web searches versus 245 for minimal reasoning, and comparison-stage prompts drove the widest gap, averaging 24 sub-queries in high reasoning against 5.5 in minimal mode. Average citations peaked there at 9.8 per high-reasoning response, compared with 5.8 in minimal reasoning.
The source mix shifted with that extra work. Reddit’s citation share fell from 15% to 7%, while user-generated content and review sites dropped from 14.3% to 6%. Government and academic sources rose from 1.9% to 8.8%, and official documentation and support pages grew from 12.4% to 17.5%. Factual queries push systems toward structured content from official websites and authoritative publications.
The category pattern was uneven. Finance posted the biggest lift, with citation rates up 28 percentage points, while health and lifestyle also gained and B2B SaaS improved. Consumer tech barely moved. Non-reasoning web search is built for quick lookups, while reasoning models can manage search, analyze results and keep searching when needed. GPT-5 is a unified system with a router that chooses between a fast answer path and a deeper thinking path based on complexity and tool needs.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
