ChatGPT vs Claude vs Gemini Statistics 2026
By Axis Intelligence Research
Co-author: Sarah Mitchell | Last updated: October 20, 2026 | License: CC BY 4.
According to Axis Intelligence Research, ChatGPT holds 58.9% of the combined ChatGPT–Gemini–Claude market on the Five-Lens Share Index (FLSI), Gemini 25.7% and Claude 15.5%, as of October 20, 2026. Claude’s share swings by 44.4 percentage points depending on the tracker, from 1.0% of the three-way AI referral split to 45.5% of three-way enterprise API spend.
Quick Answer
ChatGPT is still the largest of the three on every consumer measure, but the size of its lead depends on what is counted. Similarweb puts ChatGPT at 56.4% of generative AI web visits in September 2026, against 25.0% for Gemini and 9.6% for Claude. Statcounter’s referral data for the same month gives ChatGPT 80.48% and Claude 0.95%. In enterprise API spend, the order flips: Menlo Ventures’ survey puts Anthropic (Claude) at 40%, OpenAI at 27% and Google at 21%.
According to Axis Intelligence Research, an equal-weight blend of five independent trackers gives ChatGPT 58.9, Gemini 25.7 and Claude 15.5 points out of 100 on the Five-Lens Share Index. Drop the enterprise lens and Claude falls to 8.0, which is the clearest single measure of how much Claude’s position depends on business buyers rather than consumers.
Key Findings
- According to Axis Intelligence Research, ChatGPT scores 58.9, Gemini 25.7 and Claude 15.5 on the Five-Lens Share Index (FLSI) as of October 20, 2026, an equal-weight composite of five independent market measurements.
- According to Axis Intelligence Research, Claude’s within-three share ranges from 1.03% (Statcounter AI referrals, September 2026) to 45.45% (Menlo Ventures enterprise API spend, 2025), a 44.4-point spread on an index score of 15.5.
- ChatGPT’s share of generative AI web visits fell from 71.8% in October 2025 to 56.4% in September 2026, while Gemini’s rose from 13.6% to 25.0% and Claude’s from 2.3% to 9.6%, per Similarweb estimates.
- According to Axis Intelligence Research, ChatGPT accounts for an estimated 71.2% of the user-minutes spent across the three assistants’ mobile apps, against 54.8% of their combined monthly active users, based on Sensor Tower inputs.
- In September 2026, 43.8% of US businesses on Ramp’s platform paid for Anthropic products versus 40.6% for OpenAI, a 3.2-point gap in Anthropic’s favor, per the Ramp AI Index.
Which is bigger: ChatGPT, Gemini or Claude?
ChatGPT is bigger on four of the five lenses Axis tracks; Claude leads only in enterprise API spend. The table converts each tracker’s published shares into a “within-three share”: the assistant’s value divided by the combined value of ChatGPT, Gemini and Claude on that same tracker. That conversion strips out Copilot, Perplexity, DeepSeek and the rest, so every lens answers the same question: of the attention or money flowing to these three, who gets what?
| Measurement lens | What it counts | As of | ChatGPT | Gemini | Claude | Source |
|---|---|---|---|---|---|---|
| Web visits | Visits to genAI websites (no apps, no API) | Sep 2026 | 56.4% → 61.98 | 25.0% → 27.47 | 9.6% → 10.55 | Similarweb |
| True Audience | Deduplicated users across app and web | May 2026 | 46.4% → 54.98 | 27.7% → 32.82 | 10.3% → 12.20 | Sensor Tower (via TechCrunch) |
| AI referrals | Clicks from chatbots to websites | Sep 2026 | 80.48% → 87.16 | 10.91% → 11.82 | 0.95% → 1.03 | Statcounter |
| US adult reach | US adults who have ever used the chatbot | Feb 2026 | 44% → 59.46 | 24% → 32.43 | 6% → 8.11 | Pew Research Center |
| Enterprise API spend | Share of enterprise LLM API usage, by model vendor | Nov 2025 | 27% → 30.68 | 21% → 23.86 | 40% → 45.45 | Menlo Ventures |
| Five-Lens Share Index | Equal-weight mean of the five within-three shares | Oct 20, 2026 | 58.9 | 25.7 | 15.5 | Axis Intelligence Research |
| Consumer-only variant | Same, enterprise lens removed | Oct 20, 2026 | 65.9 | 26.1 | 8.0 | Axis Intelligence Research |
Arrow = published share → within-three share. Example for Claude on Similarweb: 9.6 ÷ (56.4 + 25.0 + 9.6) × 100 = 10.55.
ChatGPT: biggest everywhere except procurement
ChatGPT’s within-three share sits between 54.98 and 87.16 on the four consumer lenses. Its weakest reading, 30.68, comes from the enterprise lens, where Menlo Ventures attributes 27% of enterprise LLM API usage to OpenAI in 2025, down from 50% in 2023.
Gemini: the steadiest number in the set
Gemini’s five readings range from 11.82 to 32.82, a 21.0-point spread, the narrowest of the three. Its weak spot is referrals: Gemini holds 25.0% of web visits on Similarweb but 10.91% of chatbot referrals on Statcounter.
Claude: two different products in one name
Claude’s consumer-only score of 8.0 nearly doubles to 15.5 once enterprise API spend enters the index. No other assistant’s composite moves that much on a single lens.
Why do trackers disagree so much on Claude’s market share?
Because each tracker counts a different behavior, and Claude’s usage is concentrated in the behaviors most trackers miss. Statcounter counts clicks from a chatbot answer through to an external website. Claude sessions that end in code, a document or an API call produce no such click.
According to Axis Intelligence Research, Claude’s share of AI referrals is about one-tenth of its share of AI web visits (0.95% ÷ 9.6% = 0.10). For ChatGPT the same ratio is 1.43 (80.48 ÷ 56.4), and for Gemini 0.44 (10.91 ÷ 25.0). A ratio above 1 means an assistant sends out more clicks than its visit share would predict, which is what a search-like product does.
| Assistant | Similarweb web-visit share, Sep 2026 | Statcounter referral share, Sep 2026 | Referral ÷ visit ratio | Source |
|---|---|---|---|---|
| ChatGPT | 56.4% | 80.48% | 1.43 | Similarweb; Statcounter; Axis calculation |
| Gemini | 25.0% | 10.91% | 0.44 | Similarweb; Statcounter; Axis calculation |
| Claude | 9.6% | 0.95% | 0.10 | Similarweb; Statcounter; Axis calculation |
Claude is stronger in the United States than worldwide
Statcounter’s US data gives Claude 1.81% of AI chatbot referrals in September 2026, against 0.95% worldwide. According to Axis Intelligence Research, that makes Claude’s US referral share 1.91 times its global share; for Gemini the ratio is 1.22 (13.26% US vs 10.91% worldwide). Sensor Tower separately estimates Claude’s US True Audience share rose from 4.4% to nearly 14% in the 12 months to May 2026, per its State of AI 2026 summary.
Sarah Mitchell, AI/ML: The referral number is the one people screenshot, and it’s the one that tells you least about Claude. A referral is a user leaving the assistant to open a link. That is how you use a search replacement. It’s not how you use something that writes the code, drafts the contract or runs inside an agent loop through the API, and those are the jobs Claude has been bought for. So when a dashboard shows Claude below 1%, the honest reading is that Claude barely competes for the “find me a page” query, not that nobody uses it. Read the 0.10 ratio as a description of the product, not a ranking of it.
ChatGPT vs Gemini web traffic share: what changed in 12 months?
Gemini took most of the share ChatGPT lost, and then stopped gaining; Claude kept climbing. Between October 2025 and September 2026, ChatGPT’s share of generative AI web visits fell 15.4 points (71.8% to 56.4%), Gemini’s rose 11.4 points (13.6% to 25.0%) and Claude’s rose 7.3 points (2.3% to 9.6%), per Similarweb’s September 2026 update. According to Axis Intelligence Research, Claude’s web-visit share in September 2026 was 4.17 times its October 2025 level, the fastest relative gain among the seven assistants Similarweb lists.
The direction changed after April 2026. From April to September, ChatGPT gained 2.7 points (53.7% to 56.4%), Gemini lost 1.7 (26.7% to 25.0%) and Claude added 1.6 (8.0% to 9.6%).
| Month | ChatGPT | Gemini | Claude | Source |
|---|---|---|---|---|
| Oct 2025 | 71.8% | 13.6% | 2.3% | Similarweb |
| Apr 2026 | 53.7% | 26.7% | 8.0% | Similarweb |
| Jul 2026 | 54.3% | 26.6% | 9.6% | Similarweb |
| Aug 2026 | 55.6% | 25.6% | 9.3% | Similarweb |
| Sep 2026 | 56.4% | 25.0% | 9.6% | Similarweb |
Similarweb counts desktop and mobile web visits only (“All Traffic, Worldwide”); app sessions and API calls are excluded. The CSV carries all seven assistants for each month.
The rest of the field shrank over the same 12 months. Between October 2025 and September 2026, DeepSeek went from 4.1% to 3.4%, Grok from 2.8% to 2.3%, Microsoft Copilot from 2.0% to 1.5% and Perplexity from 2.5% to 0.9%, per Similarweb.
How many people use ChatGPT, Gemini and Claude?
ChatGPT has the most users on every count available, but company-reported and third-party numbers measure different things and should not be mixed. The canonical Axis figures for each assistant live on their own pages: ChatGPT users statistics, Google Gemini statistics and Claude AI statistics.
Company-reported user counts
| Assistant | Figure | Definition | As of | Source |
|---|---|---|---|---|
| ChatGPT | More than 900 million | Weekly active users, ChatGPT | Mar 31, 2026 | OpenAI (company-reported) |
| ChatGPT | 1.2 billion | “Collective” weekly users across OpenAI products | Sep 29, 2026 | OpenAI (company-reported) |
| Gemini | 950 million | Monthly active users, Gemini app | Q2 2026 | Alphabet 8-K filing, SEC |
| Claude | Not disclosed | Anthropic does not publish a consumer user count | — | — |
OpenAI’s 1.2 billion figure, given at DevDay 2026, covers OpenAI’s products together, so it is not a like-for-like update of the 900 million ChatGPT weekly users stated in OpenAI’s March 31, 2026 announcement. Alphabet’s 950 million is a monthly count, filed with the SEC for the second quarter of 2026. A weekly and a monthly active user figure cannot be ranked against each other.
Third-party estimates: users and minutes
Sensor Tower estimates more than 1.1 billion monthly active users for ChatGPT, 662 million for Gemini and 245 million for Claude at the end of May 2026, as reported by TechCrunch. Its State of AI 2026 summary puts monthly engagement at about 215 minutes per user for ChatGPT, 100 for Gemini and 120 for Claude.
According to Axis Intelligence Research, multiplying those two inputs gives ChatGPT an estimated 71.2% of the three assistants’ combined user-minutes, against 54.8% of their combined monthly users. Claude’s 120 minutes per user beat Gemini’s 100, yet both assistants lose share on this measure (Claude 12.2% of users but 8.9% of minutes) because ChatGPT users spend roughly twice as long in the app.
| Assistant | MAU (Sensor Tower, May 2026) | Minutes per user per month | Within-three MAU share | Within-three user-minute share (estimate) | Source |
|---|---|---|---|---|---|
| ChatGPT | 1,100 million (floor) | ~215 | 54.8% | 71.2% | Sensor Tower; Axis calculation |
| Gemini | 662 million | ~100 | 33.0% | 19.9% | Sensor Tower; Axis calculation |
| Claude | 245 million | ~120 | 12.2% | 8.9% | Sensor Tower; Axis calculation |
Formula: (MAU × minutes) ÷ sum over the three. Example for ChatGPT: (1,100 × 215) ÷ (1,100 × 215 + 662 × 100 + 245 × 120) = 236,500 ÷ 332,100 = 71.2%.
Across all generative AI apps, Sensor Tower projects 36 billion hours of time spent in the first half of 2026, up from 17.2 billion hours in the first half of 2025.
Who uses ChatGPT, Gemini and Claude in the United States?
ChatGPT reaches the most American adults by a wide margin, and its lead is largest among younger adults. In a February 2026 survey of 5,119 US adults, the Pew Research Center found 44% had used ChatGPT, 24% Gemini and 6% Claude.
| Chatbot | US adults who have ever used it (Feb 2026) | Source |
|---|---|---|
| ChatGPT | 44% | Pew Research Center |
| Gemini | 24% | Pew Research Center |
| Microsoft Copilot | 17% | Pew Research Center |
| Meta AI | 14% | Pew Research Center |
| Grok | 8% | Pew Research Center |
| Claude | 6% | Pew Research Center |
| Character.ai | 3% | Pew Research Center |
| Any chatbot | 49% (33% in 2024) | Pew Research Center |
ChatGPT by age
Pew found 57% of US adults under 50 had used ChatGPT, against 28% of adults 50 and older. ChatGPT’s US reach has risen in every Pew wave, from 18% in 2023 to 44% in 2026. For generational detail, see Gen Z AI statistics.
Pew’s figures date from February 2026 [older data]: it is the most recent nationally representative US survey that names Claude separately.
Claude vs ChatGPT vs Gemini for business and enterprise
Business buying is where Claude competes on even terms or better. The Ramp AI Index for October 2026 shows 43.8% of US businesses on Ramp’s platform paid Anthropic in September 2026, against 40.6% for OpenAI. OpenAI gained more over the month (+1.12 points versus +0.25), and 55.9% of Ramp businesses paid for at least one AI vendor. Ramp does not report a comparable Google figure.
| Measure | Anthropic (Claude) | OpenAI (ChatGPT/GPT) | Google (Gemini) | As of | Source |
|---|---|---|---|---|---|
| Share of Ramp businesses paying | 43.8% | 40.6% | NA | Sep 2026 | Ramp AI Index |
| Enterprise LLM API share | 40% | 27% | 21% | 2025 | Menlo Ventures |
| Enterprise LLM API share | 12% | 50% | 7% | 2023 | Menlo Ventures |
| Enterprise API share, coding | 54% | 21% | NA | 2025 | Menlo Ventures |
| API throughput | NA | More than 15 billion tokens/min (Mar 2026) | 22 billion tokens/min (Q2 2026) | 2026 | OpenAI; Alphabet (company-reported) |
Menlo Ventures, an Anthropic investor, surveyed 495 US enterprise decision-makers in November 2025 and estimated enterprise spend on foundation model APIs at $12.5 billion for 2025. Its 2025 edition is the latest available [older data]. Menlo counts model vendors, not chat apps: the Anthropic figure covers Claude models used through the API, in products that may never show the Claude name.
Google’s enterprise case rests on distribution. Alphabet says Gemini Enterprise is used by nearly 90% of the Fortune 100, and Google Cloud revenue reached $24.8 billion in Q2 2026, up 82% year on year, per its SEC filing.
Switching is becoming routine
Ramp reported that a record 8% of its businesses changed AI model provider in a given month (three-month average to September 2026). For procurement teams, that number matters more than any single month’s share: vendor lock-in at the model layer is weak.
Sarah Mitchell, AI/ML: A 3.2-point Ramp lead and a 40% Menlo share say the same thing from two directions: companies pay for Claude when the work is code or long documents, and they keep paying OpenAI for everything else. The 8% monthly switching rate is the number to watch. Model leadership changes every release cycle, and procurement now follows the eval results instead of the brand. That’s a market where today’s 3.2-point lead is one model release away from reversing.
For model-level performance rather than market share, see our hands-on coding comparison of Claude, ChatGPT and Gemini and the wider LLM statistics.
Which converts more users into paying customers: ChatGPT, Claude or Gemini?
ChatGPT has the largest paying base in absolute terms; Claude converts a larger share of its users. OpenAI said in March 2026 it had more than 50 million subscribers and more than 900 million weekly users, and was generating $2 billion in revenue per month, with enterprise above 40% of revenue.
According to Axis Intelligence Research, that puts ChatGPT’s paid subscribers at about 5.6% of its weekly users (50 ÷ 900). Sensor Tower data reported by TechCrunch puts the paying share of Anthropic’s consumer users at 13%, the highest in the field. The two figures use different denominators (weekly users versus Sensor Tower’s audience estimate), so we publish them side by side rather than as a ratio.
Revenue: what the companies themselves report
Anthropic says its run-rate revenue went from approximately $9 billion at the end of 2025 to $14 billion in February 2026, more than $30 billion in April and more than $47 billion in May 2026, per its Series G, April 2026 compute and Series H announcements. According to Axis Intelligence Research, that is a 5.2-fold increase in about five months (47 ÷ 9). Run-rate revenue is the current month’s revenue multiplied by 12.
OpenAI’s latest company figure is $2 billion of revenue per month, stated on March 31, 2026, which Axis Intelligence Research annualizes to $24 billion (2 × 12). Set against Anthropic’s own disclosure a week later (more than $30 billion on April 6, 2026), the two company-reported figures put Anthropic ahead on annualized revenue in early April 2026. OpenAI has not published a newer company figure; later press reports put it higher, so this comparison holds for that date only. Alphabet does not break out Gemini revenue.
| Company | Figure | Basis | As of | Source |
|---|---|---|---|---|
| Anthropic (Claude) | ~$9B | Run-rate revenue | End of 2025 | Anthropic (company-reported) |
| Anthropic (Claude) | $14B | Run-rate revenue | Feb 12, 2026 | Anthropic (company-reported) |
| Anthropic (Claude) | More than $30B | Run-rate revenue | Apr 6, 2026 | Anthropic (company-reported) |
| Anthropic (Claude) | More than $47B | Run-rate revenue | May 2026 | Anthropic (company-reported) |
| OpenAI (ChatGPT) | $2B per month → $24B annualized | Monthly revenue × 12 | Mar 31, 2026 | OpenAI (company-reported); Axis calculation |
| Google (Gemini) | Not disclosed | — | — | — |
Anthropic raised $65 billion in its Series H at a $965 billion post-money valuation, announced May 28, 2026.
Ads
ChatGPT is also building an ad business. About 17% of ChatGPT daily users were being served ads by May 2026, per Sensor Tower data reported by TechCrunch. Sensor Tower adds that US generative AI ad spend more than tripled in Q1 2026, led by OpenAI (+800% year on year) and Anthropic (+1,184%). Company-level financials are tracked on OpenAI statistics and Anthropic statistics.
ChatGPT vs Claude vs Gemini pricing (October 2026)
The standard paid tier costs about $20 a month for all three; the differences sit above and below it. Claude is the only one of the three without a paid tier under $10. Prices below are US list prices before tax, checked on each vendor’s own page on October 20, 2026; the ChatGPT Go price is the US price OpenAI stated at its worldwide launch on January 16, 2026.
| Tier | ChatGPT | Claude | Gemini (Google AI) | Source |
|---|---|---|---|---|
| Low-cost tier | Go $8 | — | Google AI Plus $4.99 | OpenAI; Google |
| Standard paid tier | Plus $20 | Pro $20 monthly, $17/month billed annually | Google AI Pro $19.99 | OpenAI; Anthropic; Google |
| Power-user tiers | Pro 100 $100 · Pro 200 $200 · Pro 500 $500 | Max from $100 | AI Ultra $99.99 · $199.99 | OpenAI; Anthropic; Google |
Gemini’s paid plans bundle Google storage and Workspace features, so the sticker price is not a like-for-like model comparison. Per-token API rates are covered on LLM API pricing.
Methodology
Collection. Axis Intelligence Research fetched each input from its producer on October 20, 2026: Statcounter’s public AI chatbot market share pages (worldwide and US), Similarweb’s September 2026 genAI traffic update (its post on X, all seven assistants, five months), Sensor Tower’s State of AI 2026 summary plus its end-of-May figures as published by TechCrunch, Pew Research Center’s 2026 chatbot survey, Menlo Ventures’ 2025 enterprise report, Ramp’s AI Index, OpenAI, Anthropic and Alphabet disclosures, and the three vendors’ pricing pages.
The Five-Lens Share Index (FLSI). For each lens L, the within-three share of assistant a is:
within_share(a, L) = value(a, L) ÷ [value(ChatGPT, L) + value(Gemini, L) + value(Claude, L)] × 100
FLSI(a) = 0.20 × Σ within_share(a, L) over the five lenses.
Weights are equal (0.20) because no lens is a better proxy for “market share” than another; each measures a distinct behavior: web visiting, cross-platform audience, outbound clicks, population reach and enterprise spend. The lens spread is the highest within-three share minus the lowest. The consumer-only variant drops the Menlo lens and weights the remaining four at 0.25. The FLSI does not average raw values from different methods: it averages each tracker’s internal split among the same three assistants, which is why the inputs can be combined while their absolute levels cannot. The October 20, 2026 reading is the baseline reading.
Vendor mapping. Menlo Ventures and Ramp report vendors (Anthropic, OpenAI, Google), not apps. We map them to Claude, ChatGPT and Gemini because each vendor’s flagship model family carries that name; the enterprise lens therefore includes API use that never touches the consumer app.
Other Axis calculations. User-minute share = MAU × minutes per user, normalized across the three. Referral-to-visit ratio = Statcounter share ÷ Similarweb share for the same month. US-to-world ratio = Statcounter US share ÷ worldwide share. Each formula and its result is a row in the CSV with axis_calculated = yes.
What the inputs miss. The lenses have different dates (November 2025 to September 2026), so the FLSI is a snapshot of the latest published reading of each, not a single-month measurement. Similarweb excludes apps; Statcounter sees only clicks to sites running its tracker; Pew covers US adults only; Menlo’s respondents are US enterprises and Menlo is an Anthropic investor. ChatGPT’s 1.1 billion Sensor Tower MAU is a floor (“more than”), which slightly understates its user-minute share.
About this dataset
The file chatgpt-vs-claude-vs-gemini-statistics.csv holds 192 rows in long format: one row per observation, with metric, value, unit, as_of_date, assistant, geography and segment, plus provenance columns (source_org, source_document, source_url, retrieved_date, is_primary, axis_calculated, method_note). It covers October 2025 to October 2026, with the 2023 and 2025 Menlo readings and the 2023–2026 Pew series as history. Every FLSI component, index reading and derived ratio is reproducible from the rows above it. Values are raw numbers; percentages are stored without the % sign. License: CC BY 4.0. Mirrors on Hugging Face, Kaggle and GitHub carry the same file name: [links added at upload].
How to cite this page
APA: Axis Intelligence Research, & Mitchell, S. (2026, October 20). ChatGPT vs Claude vs Gemini statistics 2026: Five trackers, three different leaders. Axis Intelligence. https://axis-intelligence.com/chatgpt-vs-claude-vs-gemini-statistics/
MLA: Axis Intelligence Research, and Sarah Mitchell. “ChatGPT vs Claude vs Gemini Statistics 2026: Five Trackers, Three Different Leaders.” Axis Intelligence, 20 Oct. 2026, axis-intelligence.com/chatgpt-vs-claude-vs-gemini-statistics/.
Chicago: Axis Intelligence Research, and Sarah Mitchell. “ChatGPT vs Claude vs Gemini Statistics 2026: Five Trackers, Three Different Leaders.” Axis Intelligence, October 20, 2026. https://axis-intelligence.com/chatgpt-vs-claude-vs-gemini-statistics/.
Dataset citation line: Axis Intelligence Research, ChatGPT vs Claude vs Gemini Statistics 2026, 2026. CC BY 4.0.
FAQ
Is Gemini catching up with ChatGPT in 2026?
It closed much of the gap between October 2025 and April 2026, then stalled. Gemini’s share of genAI web visits rose from 13.6% in October 2025 to 26.7% in April 2026, and slipped to 25.0% by September 2026, per Similarweb. Over the same April–September window ChatGPT recovered 2.7 points.
Why does Claude show under 1% market share on some charts?
Those charts usually use referral data. Statcounter measured Claude at 0.95% of worldwide AI chatbot referrals in September 2026, but Claude held 9.6% of genAI web visits the same month, per Similarweb. Claude users rarely click out to websites, so referral trackers undercount it.
Do more businesses pay for Claude or ChatGPT?
Slightly more pay for Claude, on Ramp’s data. In September 2026, 43.8% of US businesses on Ramp paid Anthropic and 40.6% paid OpenAI. OpenAI grew faster that month, adding 1.12 points to Anthropic’s 0.25.
Which has more monthly users, Gemini or ChatGPT?
ChatGPT, on Sensor Tower’s comparable estimates: more than 1.1 billion monthly active users at the end of May 2026, against 662 million for Gemini. Alphabet itself reports 950 million Gemini app monthly users for Q2 2026, a company count that is not comparable to Sensor Tower’s method.
Does anyone spend more time in Claude than in Gemini?
Yes, per user. Sensor Tower estimates about 120 minutes per user per month for Claude and 100 for Gemini, both well behind ChatGPT’s roughly 215. On total user-minutes, Gemini still ranks second because its audience is larger.
How many Americans have used Claude?
About 6% of US adults, according to Pew Research Center’s February 2026 survey of 5,119 adults. The same survey put ChatGPT at 44% and Gemini at 24%.
Is the Five-Lens Share Index the same as market share?
No. The FLSI is an Axis Intelligence Research composite that averages each assistant’s share of the three-way total across five trackers, with equal weights. It reads ChatGPT 58.9, Gemini 25.7 and Claude 15.5 as of October 20, 2026, and is meant to show agreement and disagreement between measurements, not revenue share.
How often do companies switch between OpenAI, Anthropic and Google?
More often than ever, by Ramp’s count: a record 8% of its businesses changed AI model provider in a month (three-month average to September 2026). That churn level means business share figures can move within a single quarter.