--- title: "AI Search Statistics 2026: Sourced Numbers on GEO | Reneka Digital" description: "Fifteen statistics on how ChatGPT, Perplexity, Gemini, and AI Overviews retrieve and cite sources, each linked to its primary source and dated. Maintained by Reneka Digital, Houston." url: https://renekadigital.com/resources/ai-search-statistics/ publisher: Reneka Digital, Houston, Texas --- Reference · Updated September 2026 # AI search statistics, with sources. The numbers Reneka uses when it explains generative engine optimization, each one linked to the study or vendor document it came from and dated. Nothing on this page is estimated or rounded beyond what the source reports. Figures are removed when a source is retracted or superseded. ## Fifteen statistics on AI search, September 2026. | Figure | What it measures | Detail | Source and date | |---|---|---|---| | **40%** | Maximum visibility gain from on-page edits in the only peer-reviewed GEO benchmark | Adding sourced statistics, quotations, inline citations, and clearer prose; keyword stuffing underperformed | [Aggarwal et al., GEO: Generative Engine Optimization, KDD 2024](https://arxiv.org/abs/2311.09735) (2024) | | **87%** | Share of ChatGPT search citations that matched Bing's top results | Against 56% for Google; 500+ citations across 100 queries | [Seer Interactive, February 2025](https://www.seerinteractive.com/insights/87-percent-of-searchgpt-citations-match-bings-top-results) (Feb 2025) | | **0.664** | Correlation between brand web mentions and AI Overview visibility | Backlinks: 0.218. Branded anchors: 0.527. 75,000 brands | [Ahrefs, brand correlation study of 75,000 brands, May 2025](https://ahrefs.com/blog/ai-overview-brand-correlation/) (May 2025) | | **25.7%** | How much fresher AI-cited pages are than organic results | 1,064 days versus 1,432; 16.975 million cited URLs across five AI surfaces | [Ahrefs, Do AI assistants prefer to cite fresh content, July 2025](https://ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content) (Jul 2025) | | **458 days** | How much newer ChatGPT's citations are than organic results, on average | Google AI Overviews showed almost no freshness preference (16 days older) | [Ahrefs, Do AI assistants prefer to cite fresh content, July 2025](https://ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content) (Jul 2025) | | **43.8%** | Share of pages ChatGPT cited that were "best X" lists | 26,283 source URLs behind 750 prompts; 79.1% of cited lists updated within the year | [Ahrefs, research on 'best' lists cited by ChatGPT, December 2025](https://ahrefs.com/blog/best-lists-research/) (Dec 2025) | | **35%** | Share of cited "best" lists published on low-authority domains | Same Ahrefs study; a quality problem engines have not solved | [Ahrefs, research on 'best' lists cited by ChatGPT, December 2025](https://ahrefs.com/blog/best-lists-research/) (Dec 2025) | | **61.7%** | Share of AI citations that never name the brand (ghost citations) | 3,981 domain appearances, 115 prompts, 14 countries, four engines | [Semrush and Growth Memo, The ghost citations study, June 2026](https://www.semrush.com/blog/the-ghost-citations-study/) (Jun 2026) | | **87% vs 20.7%** | ChatGPT's citation rate versus its brand mention rate | Gemini is the mirror image: 83.7% mention rate, 21.4% citation rate | [Semrush and Growth Memo, The ghost citations study, June 2026](https://www.semrush.com/blog/the-ghost-citations-study/) (Jun 2026) | | **0** | Major crawlers from OpenAI, Anthropic, Meta, ByteDance, or Perplexity that render JavaScript | Only Gemini (Googlebot infrastructure) and Applebot render; hundreds of millions of requests analyzed | [Vercel and MERJ, The rise of the AI crawler, December 2024](https://vercel.com/blog/the-rise-of-the-ai-crawler) (Dec 2024) | | **34.82%** | Share of ChatGPT crawler fetches that hit 404 pages | Plus 14.36% spent following redirects | [Vercel and MERJ, The rise of the AI crawler, December 2024](https://vercel.com/blog/the-rise-of-the-ai-crawler) (Dec 2024) | | **569 million** | GPTBot requests per month across Vercel's network | Googlebot: 4.5 billion. Claude: 370 million. PerplexityBot: 24.4 million | [Vercel and MERJ, The rise of the AI crawler, December 2024](https://vercel.com/blog/the-rise-of-the-ai-crawler) (Dec 2024) | | **97%** | Share of llms.txt files that received zero requests in May 2026 | 137,000 domains; about 1.1% of the few requests came from retrieval bots | [Ahrefs, llms.txt study of 137,000 domains, June 2026](https://ahrefs.com/blog/llmstxt-study/) (Jun 2026) | | **+2.2%** | Change in ChatGPT citations after adding JSON-LD schema | 1,885 treated pages vs 4,000 controls; AI Overviews -4.6%, AI Mode +2.4%; no major uplift on any platform | [Ahrefs, schema markup and AI citations, May 2026](https://ahrefs.com/blog/schema-ai-citations/) (May 2026) | | **24 hours** | Time OpenAI says changes to OAI-SearchBot access take to apply | Sites opted out of OAI-SearchBot are not shown in ChatGPT search answers | [OpenAI, Overview of OpenAI crawlers and fetchers](https://developers.openai.com/api/docs/bots) (2026) | ## Three cautions. - **Correlation studies describe visibility, not causation.** The mentions and freshness numbers show what cited pages have in common. They do not prove that adding a mention or a date causes a citation. - **Answer engines are non-deterministic.** The same prompt returns different sources on different runs, so any single-run "rank" is noise. Studies with large samples across many runs are the ones worth citing. - **Vendor documentation changes.** Crawler names, robots.txt behavior, and access rules are taken from the vendors' own pages as read on the date shown. Check the linked page before relying on it. ## Cite it, or ask for the current version. This page may be quoted with attribution to Reneka Digital and a link. Reneka reviews it quarterly and adds figures from new studies with public methodology. For the measurement approach behind Reneka's own client numbers, see the [methodology page](https://renekadigital.com/resources/methodology/). For definitions, see the [AI search glossary](https://renekadigital.com/resources/glossary/). ## Common questions. ### Where do these AI search statistics come from? Each row links to its primary source: the KDD 2024 GEO paper, Ahrefs, Seer Interactive, Semrush with Growth Memo, Vercel with MERJ, and vendor documentation from OpenAI. Dates are the publication or reading date shown in the table. ### How often is this page updated? Quarterly, and whenever a cited source is retracted or superseded. The visible updated date on this page and the dateModified in its schema change together. --- Source: https://renekadigital.com/resources/ai-search-statistics/ Reneka Digital, Houston, TX 77002. info@renekadigital.com. (346) 224-6908.