OpenAI was named in 258 of 288 four-engine answers (89.6%) and Google in 248 of 288 (86.1%), with overlapping intervals. OpenAI was named first in 87 answers. ChatGPT named OpenAI in 71 of 72 answers. AI Overviews appeared for all 72 lookups, and company sites supplied 83% of citations.
RESEARCH · Published August 11, 2026 · Updated October 30, 2026
AI Companies in AI Answers: Who ChatGPT, Claude, Gemini and Perplexity Name
Published August 11, 2026 · Updated October 30, 2026
288
AI answers analyzed across ChatGPT, Claude, Gemini and Perplexity
89.6%
of answers named OpenAI, the most-named (258 of 288)
50
tracked names named at least once (of 50)
100%
of Google searches showed an AI Overview (72 of 72)
Version 2. This version replaces the first version of this study, first published August 11, 2026. It uses measured counts, so the two versions are not comparable. Version 2 asks 24 questions three times each on ChatGPT, Claude, Gemini and Perplexity (288 answers) and adds Google AI Overviews. Every rate comes with its count and a 95% interval.
What the data shows
01
89.6% vs 86.1%
OpenAI was named in 258 of 288 four-engine answers (CI 85.5–92.6%) and Google in 248 of 288 (CI 81.6–89.6%). The intervals overlap, so the two are statistically tied at the top.
02
87/288 (30.2%)
OpenAI was the first company named in 87 of 288 answers. Anthropic followed with 56 of 288, Microsoft with 30 and Google with 29. Google's high mention rate comes with a later average first position (3.4, against 2.7 for OpenAI).
03
66% vs 61.5%
Anthropic (190 of 288, CI 60.3–71.2%) and Microsoft (177 of 288, CI 55.7–66.9%) overlap and cannot be ranked against each other. Meta (156 of 288, 54.2%) overlaps Microsoft.
04
71/72 (98.6%)
ChatGPT named OpenAI in 71 of 72 answers (98.6%, CI 92.5–99.8%), compared with 187 of 216 answers (86.6%) from the other three engines. It named OpenAI first in 35 of 72 answers (48.6%).
05
40/72 (55.6%)
Claude named DeepSeek in 40 of 72 answers (55.6%), against 11 of 72 (15.3%) on ChatGPT. Claude also named Moonshot AI in 26 of 72 answers, compared with 6 of 72 on ChatGPT and 5 of 72 on Gemini.
06
72/72 (100%)
A Google AI Overview appeared for all 72 lookups (100%, CI 94.9–100%). Inside the overviews, Google was named in 62 of 72 (86.1%) and OpenAI in 58 of 72 (80.6%).
07
83%
Of 2,545 four-engine citations, company and brand sites supplied 1,249 (49.1%), directories and rankings 390 (15.3%) and community and social sites 235 (9.2%).
08
26/72 (36.1%)
youtube.com was cited in 26 of 72 AI Overviews (36.1%) and reddit.com in 16 of 72 (22.2%). Community and social sites made up 33.5% of overview citations, against 4.8% for the four engines. Of 209 Google AI Overview citations, company and brand sites supplied 89 (42.6%) and community and social sites 77 (36.8%).
09
10/12 (83.3%)
Anthropic was named first in 10 of 12 answers about the best AI model for coding. OpenAI was named first in all 12 answers about AI assistants for small businesses.
Appearance rate across 288 four-engine answers. The bar is the 95% confidence interval; overlapping bars cannot be told apart at this sample size.
Which names AI engines name
OpenAI
89.6%
Google
86.1%
Anthropic
66.0%
Microsoft
61.5%
Meta
54.2%
Mistral AI
45.5%
Alibaba
39.6%
DeepSeek
39.2%
Amazon
33.0%
Scale AI
28.8%
xAI
27.4%
IBM
21.2%
Moonshot AI
17.0%
NVIDIA
17.0%
Cohere
16.3%
| # | Name | Named in | Appearance rate | 95% interval |
|---|---|---|---|---|
| 1 | OpenAI | 258/288 | 89.6% | 85.5–92.6% |
| 2 | 248/288 | 86.1% | 81.6–89.6% | |
| 3 | Anthropic | 190/288 | 66.0% | 60.3–71.2% |
| 4 | Microsoft | 177/288 | 61.5% | 55.7–66.9% |
| 5 | Meta | 156/288 | 54.2% | 48.4–59.8% |
| 6 | Mistral AI | 131/288 | 45.5% | 39.8–51.3% |
| 7 | Alibaba | 114/288 | 39.6% | 34.1–45.3% |
| 8 | DeepSeek | 113/288 | 39.2% | 33.8–45.0% |
| 9 | Amazon | 95/288 | 33.0% | 27.8–38.6% |
| 10 | Scale AI | 83/288 | 28.8% | 23.9–34.3% |
| 11 | xAI | 79/288 | 27.4% | 22.6–32.9% |
| 12 | IBM | 61/288 | 21.2% | 16.9–26.3% |
| 13 | Moonshot AI | 49/288 | 17.0% | 13.1–21.8% |
| 14 | NVIDIA | 49/288 | 17.0% | 13.1–21.8% |
| 15 | Cohere | 47/288 | 16.3% | 12.5–21.0% |
The full list
Own-domain citation is the share of 285 grounded answers that cited the entity's own domains (subdomains included). Citation share is of all 2,545 citations.
| # | Name | Named in | Appearance rate | 95% interval | First mention | Own-domain citation | Citation share |
|---|---|---|---|---|---|---|---|
| 1 | OpenAI | 258/288 | 89.6% | 85.5–92.6% | 2.7 | 5.3% (15/285) | 0.6% |
| 2 | 248/288 | 86.1% | 81.6–89.6% | 3.4 | 4.6% (13/285) | 0.6% | |
| 3 | Anthropic | 190/288 | 66% | 60.3–71.2% | 2.9 | 4.2% (12/285) | 0.6% |
| 4 | Microsoft | 177/288 | 61.5% | 55.7–66.9% | 4 | 3.5% (10/285) | 0.4% |
| 5 | Meta | 156/288 | 54.2% | 48.4–59.8% | 5.3 | 0.4% (1/285) | 0% |
| 6 | Mistral AI | 131/288 | 45.5% | 39.8–51.3% | 6.5 | 2.5% (7/285) | 0.3% |
| 7 | Alibaba | 114/288 | 39.6% | 34.1–45.3% | 5.1 | 0.4% (1/285) | 0% |
| 8 | DeepSeek | 113/288 | 39.2% | 33.8–45% | 5.6 | 0.4% (1/285) | 0% |
| 9 | Amazon | 95/288 | 33% | 27.8–38.6% | 5.6 | 3.2% (9/285) | 0.4% |
| 10 | Scale AI | 83/288 | 28.8% | 23.9–34.3% | 5.6 | 0% (0/285) | 0% |
| 11 | xAI | 79/288 | 27.4% | 22.6–32.9% | 6.3 | 0% (0/285) | 0% |
| 12 | IBM | 61/288 | 21.2% | 16.9–26.3% | 7 | 2.8% (8/285) | 0.3% |
| 13 | Moonshot AI | 49/288 | 17% | 13.1–21.8% | 4.8 | 0% (0/285) | 0% |
| 14 | NVIDIA | 49/288 | 17% | 13.1–21.8% | 7.5 | 1.4% (4/285) | 0.2% |
| 15 | Cohere | 47/288 | 16.3% | 12.5–21% | 7.4 | 2.1% (6/285) | 0.2% |
| 16 | Databricks | 41/288 | 14.2% | 10.7–18.7% | 6.3 | 2.1% (6/285) | 0.2% |
| 17 | Zhipu AI | 40/288 | 13.9% | 10.4–18.4% | 5.5 | 0.4% (1/285) | 0% |
| 18 | Salesforce | 40/288 | 13.9% | 10.4–18.4% | 6.4 | 0% (0/285) | 0% |
| 19 | Hugging Face | 36/288 | 12.5% | 9.2–16.8% | 5.8 | 7.7% (22/285) | 1.1% |
| 20 | MiniMax | 32/288 | 11.1% | 8–15.3% | 6.9 | 0% (0/285) | 0% |
| 21 | Baidu | 30/288 | 10.4% | 7.4–14.5% | 7.7 | 0.4% (1/285) | 0% |
| 22 | Tencent | 29/288 | 10.1% | 7.1–14.1% | 7.2 | 0.7% (2/285) | 0.1% |
| 23 | Stability AI | 29/288 | 10.1% | 7.1–14.1% | 5.5 | 1.1% (3/285) | 0.1% |
| 24 | Perplexity | 24/288 | 8.3% | 5.7–12.1% | 6.1 | 1.1% (3/285) | 0.1% |
| 25 | Zapier | 23/288 | 8% | 5.4–11.7% | 4.7 | 8.4% (24/285) | 0.9% |
The top eight, side by side
| # | Name | Appearance rate | Answers and interval | Other measures |
|---|---|---|---|---|
| 1 | OpenAI | 89.6% | 258/288 answers · 95% CI 85.5–92.6% | first-mention position: 2.7; owned-domain citation (15/285): 5.3%; citation share: 0.6% |
| 2 | 86.1% | 248/288 answers · 95% CI 81.6–89.6% | first-mention position: 3.4; owned-domain citation (13/285): 4.6%; citation share: 0.6% | |
| 3 | Anthropic | 66.0% | 190/288 answers · 95% CI 60.3–71.2% | first-mention position: 2.9; owned-domain citation (12/285): 4.2%; citation share: 0.6% |
| 4 | Microsoft | 61.5% | 177/288 answers · 95% CI 55.7–66.9% | first-mention position: 4; owned-domain citation (10/285): 3.5%; citation share: 0.4% |
| 5 | Meta | 54.2% | 156/288 answers · 95% CI 48.4–59.8% | first-mention position: 5.3; owned-domain citation (1/285): 0.4%; citation share: 0% |
| 6 | Mistral AI | 45.5% | 131/288 answers · 95% CI 39.8–51.3% | first-mention position: 6.5; owned-domain citation (7/285): 2.5%; citation share: 0.3% |
| 7 | Alibaba | 39.6% | 114/288 answers · 95% CI 34.1–45.3% | first-mention position: 5.1; owned-domain citation (1/285): 0.4%; citation share: 0% |
| 8 | DeepSeek | 39.2% | 113/288 answers · 95% CI 33.8–45% | first-mention position: 5.6; owned-domain citation (1/285): 0.4%; citation share: 0% |
All four engines returned 72 answers. ChatGPT, Gemini and Perplexity grounded all 72 in sources, and Claude grounded 69. ChatGPT named OpenAI in 71 of 72 answers and Microsoft in 54 of 72 (75%), compared with 35 of 72 (48.6%) on Claude. Claude named Chinese labs more often than the other engines. Zhipu AI appeared in 21 of 72 Claude answers and 4 of 72 ChatGPT answers.
ChatGPT
72 asked · 72 returned · 72 grounded
Same questions, different answers
OpenAI
98.6%
Google
80.6%
Microsoft
75.0%
Anthropic
62.5%
Meta
52.8%
Mistral AI
43.1%
Claude
72 asked · 72 returned · 69 grounded
Google
90.3%
OpenAI
87.5%
Anthropic
70.8%
DeepSeek
55.6%
Meta
50.0%
Microsoft
48.6%
Gemini
72 asked · 72 returned · 72 grounded
OpenAI
90.3%
Google
88.9%
Meta
70.8%
Anthropic
69.4%
Microsoft
68.1%
Mistral AI
56.9%
Perplexity
72 asked · 72 returned · 72 grounded
Google
84.7%
OpenAI
81.9%
Anthropic
61.1%
Microsoft
54.2%
Alibaba
50.0%
DeepSeek
44.4%
Google AI Overviews were captured through 72 U.S. desktop searches and scored only where an overview appeared, so five-engine figures use 288 + 72 = 360 answers.
The fifth engine
100%
Trigger rate: 72 of 72 searches
Google AI Overviews appeared for all 72 of 72 lookups, a trigger rate of 100% (CI 94.9–100%), and every overview was scored. Google was named in 62 of 72 overviews (86.1%), OpenAI in 58 of 72 (80.6%), Anthropic in 45 of 72 (62.5%) and Microsoft in 36 of 72 (50%). Some companies appeared much less often in overviews than in four-engine answers. Mistral AI was named in 7 of 72 overviews (9.7%), against 45.5% across the four engines. Cohere did not appear in any of the 72 overviews. Zapier appeared in 13 of 72 overviews (18.1%).
| Five-engine rank | Name | Five-engine rate | Four-engine rank |
|---|---|---|---|
| 1 | OpenAI | 87.8% | 1 |
| 2 | 86.1% | 2 | |
| 3 | Anthropic | 65.3% | 3 |
| 4 | Microsoft | 59.2% | 4 |
| 5 | Meta | 50.0% | 5 |
| 6 | Mistral AI | 38.3% | 6 |
| 7 | DeepSeek | 36.4% | 8 |
| 8 | Alibaba | 34.2% | 7 |
Who each question surfaces first
Six questions from the set. Each tile uses that question's own four-engine answers as the denominator.
| Question | Leader | Next | AI Overview |
|---|---|---|---|
| Q01 · 12 answers | OpenAI 100.0% | Anthropic 100.0%; Google 100.0% | AI Overview shown on 3 of 3 searches |
| Q02 · 12 answers | OpenAI 100.0% | Anthropic 100.0%; Google 100.0% | AI Overview shown on 3 of 3 searches |
| Q03 · 12 answers | OpenAI 100.0% | Google 100.0%; Microsoft 100.0% | AI Overview shown on 3 of 3 searches |
| Q04 · 12 answers | Mistral AI 100.0% | Meta 91.7%; DeepSeek 83.3% | AI Overview shown on 3 of 3 searches |
| Q05 · 12 answers | OpenAI 91.7% | Anthropic 91.7%; Google 91.7% | AI Overview shown on 3 of 3 searches |
| Q06 · 12 answers | OpenAI 100.0% | Anthropic 100.0%; Google 100.0% | AI Overview shown on 3 of 3 searches |
The four engines cited 2,545 sources. Of 2,545 four-engine citations, company and brand sites supplied 1,249 (49.1%), directories and rankings 390 (15.3%), community and social sites 235 (9.2%) and institutions and trade bodies 63 (2.5%). The most-cited domains were benchlm.ai (62 citations, in 42 answers), en.wikipedia.org (60, in 37 answers), artificialanalysis.ai (38, in 31 answers) and llm-stats.com (38).
Four AI engines · share of 285 grounded answers
Most-cited domains
benchlm.ai
14.7%
en.wikipedia.org
13.0%
artificialanalysis.ai
10.9%
arena.ai
9.5%
llm-stats.com
9.1%
zapier.com
8.4%
huggingface.co
7.7%
gartner.com
6.3%
arxiv.org
6.0%
github.com
5.6%
Google AI Overviews · share of 72 overviews
youtube.com
36.1%
reddit.com
22.2%
linkedin.com
11.1%
zapier.com
8.3%
facebook.com
6.9%
cloud.google.com
4.2%
coval.ai
4.2%
benchlm.ai
4.2%
geotoolbox.ai
4.2%
flaviocopes.com
4.2%
Citation share by source type
Share of all citations by the type of site cited. Source types were reclassified on October 29, 2026. 'Company / brand sites' now covers only the official sites of companies, brands and schools; publishers, directories, trade bodies and other sites have their own types, and sites that could not be placed are 'Other / unclassified'.
| Source type | Four AI engines (of 2,545 citations) | Google AI Overviews (of 209 citations) |
|---|---|---|
| Company / brand sites | 1,249 (49.1%) | 89 (42.6%) |
| Directories / rankings | 390 (15.3%) | 7 (3.3%) |
| Community / social | 235 (9.2%) | 77 (36.8%) |
| Review / comparison publishers | 190 (7.5%) | 8 (3.8%) |
| Research / data providers | 175 (6.9%) | 8 (3.8%) |
| Trade / specialist press | 96 (3.8%) | 8 (3.8%) |
| General news / business press | 70 (2.8%) | 5 (2.4%) |
| Reference | 66 (2.6%) | 1 (0.5%) |
| Institutions / trade bodies | 63 (2.5%) | 6 (2.9%) |
| Official / government | 6 (0.2%) | 0 (0%) |
| Other / unclassified | 5 (0.2%) | 0 (0%) |
How can names improve their AI visibility?
These steps follow from the measured results above. They describe what the engines cited in this study and are not a guarantee of results.
| Step | What this study measured |
|---|---|
| Put your facts on your own domain. | OpenAI's own domains were cited in 5.3% of grounded answers, Google's in 4.6% and Anthropic's in 4.2%. |
| Keep your facts accurate on the sites engines cite. | The most-cited sites other than the tracked names' own were benchlm.ai (in 14.7% of grounded answers), en.wikipedia.org (13.0%), artificialanalysis.ai (10.9%), arena.ai (9.5%) and llm-stats.com (9.1%). Directories and rankings supplied 15.3% of citations and community and social 9.2%. |
| Check each engine separately. | Claude named DeepSeek in 55.6% of answers against 15.3% on ChatGPT. Gemini named Scale AI in 47.2% of answers against 11.1% on Claude. |
| Do not ignore video and community sites. | Community and social sites supplied 36.8% of Google AI Overview citations against 9.2% in the four chat engines. youtube.com was cited in 36.1% of overviews, reddit.com in 22.2% and linkedin.com in 11.1%. |
| Re-run the same questions every month. | Compare counts and intervals, not ranks. The 24 questions are in the appendix. |
Frequently asked questions
Which AI company do AI assistants mention most often?
OpenAI and Google. OpenAI was named in 258 of 288 four-engine answers (89.6%, CI 85.5–92.6%) and Google in 248 of 288 (86.1%, CI 81.6–89.6%). The intervals overlap, so neither company can be called the leader. OpenAI was named first more often, in 87 of 288 answers against 29 for Google.
Does ChatGPT favour OpenAI?
ChatGPT named OpenAI in 71 of 72 answers (98.6%, CI 92.5–99.8%). The other three engines named OpenAI in 187 of 216 answers (86.6%). ChatGPT placed OpenAI first in 35 of 72 answers (48.6%).
Does Claude favour Anthropic?
Claude named Anthropic in 51 of 72 answers (70.8%, CI 59.5–80.1%), while the other engines named it in 139 of 216 (64.4%). The interval includes the other engines' rate. Claude named Anthropic first in 20 of 72 answers (27.8%) and did not cite Anthropic's domain in any of its 69 grounded answers.
How often do Google AI Overviews appear for questions about AI companies?
An overview appeared for all 72 of 72 lookups (100%, CI 94.9–100%). Google was named in 62 of 72 overviews and OpenAI in 58 of 72.
What sources do AI engines cite about AI companies?
Of 2,545 four-engine citations, company and brand sites supplied 1,249 (49.1%), directories and rankings 390 (15.3%) and community and social sites 235 (9.2%). AI Overviews leaned on youtube.com, which was cited in 26 of 72 overviews.
Which Chinese AI companies appear most often?
Alibaba was named in 114 of 288 answers (39.6%) and DeepSeek in 113 of 288 (39.2%). Their intervals overlap. Claude named DeepSeek in 40 of 72 answers, and ChatGPT named it in 11 of 72.
How this study was measured
24 questions × 3 runs each × 4 AI engines = 288 answers + 72 Google searches · 72 AI Overviews
24 questions, each asked 3 times on each engine, measured October 26, 2026. Market: United States, English. The destination site was not mentioned to any engine. Entities count as named when they appear in the answer text or the extracted entity list; products, imprints and research arms roll up to their parent.
| Engine | Model IDs called | Asked | Returned | Grounded |
|---|---|---|---|---|
| ChatGPT | gpt-5-mini | 72 | 72 | 72 |
| Claude | claude-sonnet-5-5 | 72 | 72 | 69 |
| Gemini | gemini-3.7-flash | 72 | 72 | 72 |
| Perplexity | sonar-pro | 72 | 72 | 72 |
| Google AI Overviews | DataForSEO Google SERP (live, advanced) | 72 | 72 shown | 72 |
Own-domain citations count a domain plus all its subdomains, with the same rule in every table.
Definitions
- Appearance rate: answers naming the entity ÷ answers returned, with Wilson 95% intervals.
- First mention: average order in which the entity first appears (1 = first).
- Own-domain citation: grounded answers citing the entity's domains ÷ grounded answers.
- Citation share: citations to the entity's domains ÷ all citations.
Limits
- Measured on one day, October 26, 2026. Answers came from each engine's API (model IDs above); the consumer apps can answer differently.
- The question set has 24 questions with 12 four-engine answers each. Questions 1–12 came from the original card and questions 13–24 were added. A different question mix would produce different rates.
- Claude returned 72 answers, but only 69 were grounded in sources. Rates based on its own-domain citations use 69 as the denominator.
- Mention rate measures whether a company is named, not whether the description is accurate or favourable.
- Names are counted when they appear in an answer's text, so longer forms such as legal company names (for example "Meta Platforms, Inc." or "Ralph Lauren Corporation") count toward the brand.
- One question, "OpenAI vs Anthropic vs Google for enterprise AI?", names three companies, so those counts include up to 12 answers to it. Without it, OpenAI and Google are still the top two.
- Intervals treat each answer as independent, so they understate uncertainty for the 24 repeated questions.
Appendix: The 24 questions
Questions marked "Core question" are the original study questions, asked word for word. The others are added neutral questions.
| # | Question | Group | Source |
|---|---|---|---|
| 01 | Which AI company makes the best large language model in 2026? | AI companies | Core question |
| 02 | What is the best AI model for coding? | AI companies | Core question |
| 03 | Which AI assistant is best for a small business? | AI companies | Core question |
| 04 | What are the best open-source large language models? | AI companies | Core question |
| 05 | Which AI companies lead on AI safety research? | AI companies | Core question |
| 06 | OpenAI vs Anthropic vs Google for enterprise AI? | AI companies | Core question |
| 07 | What is the best AI API for building a customer support chatbot? | AI companies | Core question |
| 08 | Which AI labs are leading AI research? | AI companies | Core question |
| 09 | What are the best AI models for image and video generation? | AI companies | Core question |
| 10 | Which Chinese AI labs make competitive models? | AI companies | Core question |
| 11 | What are the best enterprise AI platforms for large companies? | AI companies | Core question |
| 12 | Which AI model is best for analyzing long documents? | AI companies | Core question |
| 13 | Which companies make the most widely used AI chatbots? | AI companies | Added question |
| 14 | Which AI models are best for writing and editing text? | AI companies | Added question |
| 15 | What are the leading AI coding assistants for developers? | AI companies | Added question |
| 16 | Which AI companies offer the cheapest API pricing per token? | AI companies | Added question |
| 17 | Which AI models perform best on reasoning benchmarks? | AI companies | Added question |
| 18 | Which AI companies have raised the most funding? | AI companies | Added question |
| 19 | What are the best AI tools for data analysis? | AI companies | Added question |
| 20 | Which AI models can run locally on a laptop? | AI companies | Added question |
| 21 | Which AI companies publish the most research papers? | AI companies | Added question |
| 22 | What are the best AI voice and speech models? | AI companies | Added question |
| 23 | Which AI platforms are best for building AI agents? | AI companies | Added question |
| 24 | Which European AI companies make competitive models? | AI companies | Added question |
About 5W
5W is an independent PR, creative and GEO agency founded in 2003, and we get you in the chatbox. All of our work is done in house, including media relations, site building and content. Learn more about our GEO practice or write to info@5wpr.com.
Full report
Read the PDF
Get in touch
Let's build your next chapter.
Tell us what you're working on. A senior strategist will respond within one business day.
- info@5wpr.com
- Phone
- 212.999.5585
- Offices
- New York · HQ469 7th Avenue, Floor 8
New York, NY 10018Miami100 SE 2nd Street, Floor 38
Miami, FL 33131Tampa110 South 12th Street
Tampa, FL 33602