Which sources does Gemini cite most?
As of November 17, 2026, Reddit is the most-cited source in Gemini answers, cited in 6.87% of 7,789 observed Gemini runs.
A rate is the share of Gemini's observed runs, across every category in the window, that cited the source. Engines answer different numbers of runs, so compare rates, not counts.
Machine Relations Index, created by Jaxon Parrott. Data CC BY 4.0.
Ranking
Sources ranked by citation rate
| Rank | Source | Source role | Citation rate | Cited runs | Engine's observed runs | Confidence |
|---|---|---|---|---|---|---|
| #1 of 8,366 | Community and social platform | 6.87% | 535 | 7,789 | A | |
| #2 of 8,366 | YouTube (only pages not mapped to a channel or publication) | Search or media platform | 4.46% | 347 | 7,789 | A |
| #3 of 8,366 | Medium | Community and social platform | 3.71% | 289 | 7,789 | A |
| #4 of 8,366 | Forbes | Editorial publication | 2.34% | 182 | 7,789 | A |
| #5 of 8,366 | gartner.com | Analyst and consulting research | 1.57% | 122 | 7,789 | A |
| #6 of 8,366 | NIH | Academic and government source | 1.37% | 107 | 7,789 | A |
| #7 of 8,366 | Healthline | Editorial publication | 1.05% | 82 | 7,789 | A |
| #8 of 8,366 | G2 | Market and company database | 0.98% | 76 | 7,789 | A |
| #9 of 8,366 | ibm.com | Vendor-owned source | 0.95% | 74 | 7,789 | A |
| #10 of 8,366 | landbase.com | Vendor-owned source | 0.91% | 71 | 7,789 | A |
| #11 of 8,366 | arxiv.org | Academic and government source | 0.83% | 65 | 7,789 | A |
| #12 of 8,366 | microsoft.com | Vendor-owned source | 0.76% | 59 | 7,789 | A |
| #13 of 8,366 | github.com | Community and social platform | 0.74% | 58 | 7,789 | A |
| #14 of 8,366 | tryprofound.com | Vendor-owned source | 0.73% | 57 | 7,789 | B |
| #15 of 8,366 | dev.to | Community and social platform | 0.68% | 53 | 7,789 | A |
| #16 of 8,366 | substack.com | Community and social platform | 0.67% | 52 | 7,789 | A |
| #17 of 8,366 | dageno.ai | Other observed source | 0.64% | 50 | 7,789 | C |
| #18 of 8,366 | semrush.com | Vendor-owned source | 0.63% | 49 | 7,789 | B |
| #19 of 8,366 | tracxn.com | Market and company database | 0.60% | 47 | 7,789 | B |
| #20 of 8,366 | hubspot.com | Vendor-owned source | 0.58% | 45 | 7,789 | A |
| #21 of 8,366 | sentinelone.com | Vendor-owned source | 0.56% | 44 | 7,789 | B |
| #22 of 8,366 | amazon.com | Other observed source | 0.53% | 41 | 7,789 | A |
| #22 of 8,366 | paloaltonetworks.com | Vendor-owned source | 0.53% | 41 | 7,789 | A |
| #22 of 8,366 | shadow.inc | Other observed source | 0.53% | 41 | 7,789 | B |
| #25 of 8,366 | seranking.com | Other observed source | 0.50% | 39 | 7,789 | B |
| #25 of 8,366 | spheron.network | Other observed source | 0.50% | 39 | 7,789 | B |
| #27 of 8,366 | wealthfront.com | Other observed source | 0.49% | 38 | 7,789 | C |
| #27 of 8,366 | zapier.com | Vendor-owned source | 0.49% | 38 | 7,789 | A |
| #29 of 8,366 | databricks.com | Vendor-owned source | 0.48% | 37 | 7,789 | B |
| #30 of 8,366 | bankrate.com | Other observed source | 0.46% | 36 | 7,789 | B |
| #30 of 8,366 | Business Insider | Editorial publication | 0.46% | 36 | 7,789 | B |
| #30 of 8,366 | harvard.edu | Academic and government source | 0.46% | 36 | 7,789 | B |
| #33 of 8,366 | improvado.io | Vendor-owned source | 0.44% | 34 | 7,789 | B |
| #33 of 8,366 | nvidia.com | Vendor-owned source | 0.44% | 34 | 7,789 | B |
| #35 of 8,366 | grandviewresearch.com | Market and company database | 0.42% | 33 | 7,789 | B |
| #36 of 8,366 | iherb.com | Other observed source | 0.41% | 32 | 7,789 | C |
| #36 of 8,366 | mordorintelligence.com | Market and company database | 0.41% | 32 | 7,789 | B |
| #36 of 8,366 | otterly.ai | Other observed source | 0.41% | 32 | 7,789 | B |
| #39 of 8,366 | betterment.com | Other observed source | 0.40% | 31 | 7,789 | C |
| #39 of 8,366 | crunchbase.com | Market and company database | 0.40% | 31 | 7,789 | B |
| #39 of 8,366 | salesforce.com | Vendor-owned source | 0.40% | 31 | 7,789 | B |
| #42 of 8,366 | guideflow.com | Other observed source | 0.39% | 30 | 7,789 | A |
| #42 of 8,366 | llmpulse.ai | Other observed source | 0.39% | 30 | 7,789 | B |
| #42 of 8,366 | target.com | Other observed source | 0.39% | 30 | 7,789 | B |
| #42 of 8,366 | thorne.com | Other observed source | 0.39% | 30 | 7,789 | C |
| #46 of 8,366 | AIMultiple | Editorial publication | 0.37% | 29 | 7,789 | B |
| #46 of 8,366 | researchgate.net | Academic and government source | 0.37% | 29 | 7,789 | B |
| #46 of 8,366 | thefeed.com | Other observed source | 0.37% | 29 | 7,789 | C |
| #46 of 8,366 | Wikipedia | Academic and government source | 0.37% | 29 | 7,789 | A |
| #50 of 8,366 | experian.com | Vendor-owned source | 0.36% | 28 | 7,789 | B |
| #50 of 8,366 | failory.com | Market and company database | 0.36% | 28 | 7,789 | B |
| #50 of 8,366 | truefoundry.com | Vendor-owned source | 0.36% | 28 | 7,789 | B |
| #50 of 8,366 | yotpo.com | Other observed source | 0.36% | 28 | 7,789 | C |
| #54 of 8,366 | stripe.com | Vendor-owned source | 0.35% | 27 | 7,789 | B |
| #55 of 8,366 | beomniscient.com | Other observed source | 0.33% | 26 | 7,789 | C |
| #55 of 8,366 | clevelandclinic.org | Academic and government source | 0.33% | 26 | 7,789 | B |
| #55 of 8,366 | deloitte.com | Analyst and consulting research | 0.33% | 26 | 7,789 | B |
| #55 of 8,366 | redhat.com | Vendor-owned source | 0.33% | 26 | 7,789 | B |
| #55 of 8,366 | siftly.ai | Other observed source | 0.33% | 26 | 7,789 | C |
| #55 of 8,366 | unbiased.com | Vendor-owned source | 0.33% | 26 | 7,789 | C |
| #55 of 8,366 | WebMD | Editorial publication | 0.33% | 26 | 7,789 | B |
| #62 of 8,366 | discoveredlabs.com | Other observed source | 0.32% | 25 | 7,789 | C |
| #62 of 8,366 | fidelity.com | Other observed source | 0.32% | 25 | 7,789 | B |
| #62 of 8,366 | quora.com | Community and social platform | 0.32% | 25 | 7,789 | B |
| #65 of 8,366 | innerbody.com | Other observed source | 0.31% | 24 | 7,789 | C |
| #65 of 8,366 | MarTech | Editorial publication | 0.31% | 24 | 7,789 | B |
| #65 of 8,366 | naturemade.com | Other observed source | 0.31% | 24 | 7,789 | C |
| #65 of 8,366 | nicklafferty.com | Other observed source | 0.31% | 24 | 7,789 | C |
| #65 of 8,366 | PR Newswire | Wire and press-release distribution | 0.31% | 24 | 7,789 | B |
| #70 of 8,366 | agilitypr.com | Vendor-owned source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | facebook.com | Community and social platform | 0.30% | 23 | 7,789 | A |
| #70 of 8,366 | firecrawl.dev | Other observed source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | fortunebusinessinsights.com | Market and company database | 0.30% | 23 | 7,789 | B |
| #70 of 8,366 | frase.io | Other observed source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | geoptie.com | Other observed source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | owasp.org | Academic and government source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | People Managing People | Editorial publication | 0.30% | 23 | 7,789 | A |
| #70 of 8,366 | ringcentral.com | Other observed source | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | Kitchn | Editorial publication | 0.30% | 23 | 7,789 | C |
| #70 of 8,366 | walmart.com | Other observed source | 0.30% | 23 | 7,789 | B |
| #70 of 8,366 | Y Combinator | Community and social platform | 0.30% | 23 | 7,789 | B |
| #82 of 8,366 | CNET | Editorial publication | 0.28% | 22 | 7,789 | B |
| #82 of 8,366 | useomnia.com | Other observed source | 0.28% | 22 | 7,789 | C |
| #82 of 8,366 | vanguard.com | Other observed source | 0.28% | 22 | 7,789 | C |
| #82 of 8,366 | xseek.io | Other observed source | 0.28% | 22 | 7,789 | C |
| #86 of 8,366 | angi.com | Other observed source | 0.27% | 21 | 7,789 | B |
| #86 of 8,366 | bbcgoodfood.com | Other observed source | 0.27% | 21 | 7,789 | C |
| #86 of 8,366 | CRN | Editorial publication | 0.27% | 21 | 7,789 | C |
| #86 of 8,366 | ellty.com | Vendor-owned source | 0.27% | 21 | 7,789 | C |
| #86 of 8,366 | marketsandmarkets.com | Market and company database | 0.27% | 21 | 7,789 | B |
| #86 of 8,366 | mayoclinic.org | Other observed source | 0.27% | 21 | 7,789 | B |
| #86 of 8,366 | oracle.com | Vendor-owned source | 0.27% | 21 | 7,789 | B |
| #86 of 8,366 | ubiehealth.com | Other observed source | 0.27% | 21 | 7,789 | B |
| #86 of 8,366 | zenml.io | Other observed source | 0.27% | 21 | 7,789 | C |
| #95 of 8,366 | cisco.com | Other observed source | 0.26% | 20 | 7,789 | B |
| #95 of 8,366 | everydayhealth.com | Editorial publication | 0.26% | 20 | 7,789 | B |
| #95 of 8,366 | explodingtopics.com | Other observed source | 0.26% | 20 | 7,789 | C |
| #95 of 8,366 | marcus.com | Other observed source | 0.26% | 20 | 7,789 | C |
| #95 of 8,366 | Money | Editorial publication | 0.26% | 20 | 7,789 | C |
| #95 of 8,366 | petmd.com | Other observed source | 0.26% | 20 | 7,789 | B |
Showing the first 100 of 8,366 ranked sources.
Categories
Whole-window rates among published segments
Each link ranks sources with a published segment in that category by this engine's citation rate across the whole window. A collecting category has no published segment, so it publishes no ranking.- AI Infrastructure
- AI Security & Privacy
- AI Visibility & GEO
- Consumer Finance
- Consumer Health
- Consumer Products
- Cybersecurity
- Deep Tech & Hardware
- Education & Training
- Emergent Prosumer
- Enterprise Software
- Family Software
- Fintech
- Healthcare Services
- HR & Talent
- iGaming & Betting
- Industrial
- Legacy News Topics
- Local Services
- Logistics & Freight
- Martech & Advertising
- Physical Consumer Products
- Professional Services
- Sales, GTM & CRM
- VC & Private Equity
Method
What each engine's figures measure
Published November 17, 2026. The observation window runs May 10, 2026 to November 17, 2026. Methodology v2.1.
The latest-successful model recorded for Gemini is Gemini 3.8 Flash; it does not identify every run in the window.
Gemini's sample is 7,789 observed runs across 158 dates. The evidence floor is 10 observed runs and 7 distinct dates.
Models across the window. Whole-window engine figures pool the models used during that window. The resolved model is latest-success metadata, not attribution for every historical run. Model periods in the release manifest and engine responses count successful, public-eligible returned runs before engine-day denominator exclusions, not category-specific runs or citation counts. A null model means not recorded; missing or null periods mean historical attribution is unavailable. First and last dates do not imply uninterrupted use. A model change alone is not a collection regime.
Citation coverage and collection. Roster status includes citation yield. Degraded or quiet does not alone prove failed collection: responses can return without citations.
ChatGPT runs. They go through OpenAI's API with web search switched on. The model can change during the window; the release's model periods preserve that history, including runs whose model was not recorded. That is the retrieval layer apps and research agents built on OpenAI's API use. The signed-in ChatGPT app answers with its own models and search pipeline and cites differently: a preprint submitted on 2026-10-11 and not yet peer reviewed found that Google Search and OpenAI's API on comparable models together recovered about half of the hostnames ChatGPT Web cited (48.7 and 50.3 percent on two model settings) and 26 to 28 percent of its exact URLs (arXiv 2610.15382v1). Claude, Gemini and Perplexity runs are likewise API runs with web search; their historical models are recorded in the release's model periods when known.
Method note, 2026-09-25. From that date, Google AI Mode and Google AI Overviews runs come through a different collection path than earlier runs, with no overlapping days. Google AI Overviews answers collected on the two paths differ in how many distinct domains each cites, so its per-answer figures are read within one side of 2026-09-25, not as a trend across it. Google AI Mode changed earlier: its answers began citing far fewer distinct domains between 2026-08-10 and 2026-08-30, on the earlier path, and a same-hour comparison of the two paths on 2026-11-05 found both citing few distinct domains per answer. Google AI Mode figures from 2026-08-31 on are one collection regime, reported on their own, and are not compared or trended across the August change.
Degraded coverage: 5 of 6 answer engines are currently returning citation-bearing answers. Gemini is quiet (2 of 91 collected calls produced an observation, 2%). Coverage across the window is uneven: Gemini returned citation-bearing answers on 158 of 185 observed days (2,342 of 8,324 returned runs produced a citation); Google AI Mode returned citation-bearing answers on 142 of 185 observed days (1,920 of 2,466 returned runs produced a citation); Google AI Overviews returned citation-bearing answers on 125 of 185 observed days (1,123 of 1,659 returned runs produced a citation). Engines below 90% of the window are named here; per-engine day and run counts for all 6 are in the engine roster.
Cite this
Machine Relations Index, created by Jaxon Parrott — machinerelations.ai. As of November 17, 2026, Reddit is the most-cited source in Gemini answers, cited in 6.87% of 7,789 observed Gemini runs. Release mri_score_v2.1+2026-11-17+e203846adf5d. CC BY 4.0. https://machinerelations.ai/index/engines/geminiParrott, J. (2026). Which sources does Gemini cite most? (Version mri_score_v2.1+2026-11-17+e203846adf5d) [Data set]. Machine Relations Research. https://machinerelations.ai/index/engines/gemini https://doi.org/10.5281/zenodo.23652423@misc{parrott2026mrigemini,
author = {Parrott, Jaxon},
title = {Which sources does Gemini cite most?},
year = {2026},
version = {mri\_score\_v2.1+2026-11-17+e203846adf5d},
publisher = {Machine Relations Research},
url = {https://machinerelations.ai/index/engines/gemini},
doi = {10.5281/zenodo.23652423},
license = {CC-BY-4.0},
note = {Machine Relations Index, created by Jaxon Parrott — machinerelations.ai}
}