October 2026 AI chatbot market share data: usage vs referral vs B2B referral share, and how to weight your AI visibility score.
AI chatbot market share depends on which ruler you use. StatCounter's September 2026 referral data gives ChatGPT 80.48% of the clicks AI chatbots send to websites. Similarweb's August 2026 visit data gives ChatGPT 55.5% of the traffic AI chatbots receive. Claude holds 9.3% of visits and 0.95% of referrals. Perplexity holds 0.9% of visits and 5.85% of referrals. If your AI visibility report weights every engine the same, these gaps distort the number your team reports upward.
This week's data update covers three lenses: usage (who people visit), referral (who sends clicks out), and B2B referral (who sends clicks to business sites). Each lens ranks the engines differently. Here's what changed, why the sources disagree, and how to weight your AI engine tracking for Q4 2026.
ChatGPT leads every AI chatbot market share dataset, but its margin swings from 30 to 70 points depending on the measure. The table below lines up the three most recent public readings.
| Engine | Web visits (Similarweb, Aug 2026) | Worldwide referrals (StatCounter, Sep 2026) | B2B referrals (Octane11, Jun 2026) |
|---|---|---|---|
| ChatGPT | 55.5% | 80.48% | ~79% |
| Gemini | 25.6% | 10.91% | 9.8% |
| Claude | 9.3% | 0.95% | 4.5% |
| Perplexity | 0.9% | 5.85% | 4.2% |
| Copilot | 1.6% | 1.79% | 2.5% |
| DeepSeek | 3.4% | 0.02% | n/a |
StatCounter bases its figures on more than 3 billion monthly page views across its tracked sites. Similarweb measures visits to the chatbot websites themselves and excludes mobile apps and desktop clients. Octane11 measured 400 million sessions on B2B websites between September 2025 and June 2026. Each method produces its own leaderboard.
The two sources measure opposite ends of the same pipe. Similarweb counts people arriving at chatgpt.com, gemini.google.com, or claude.ai. StatCounter counts people leaving those products and landing on a website that runs its tracking code. One measures attention. The other measures outbound clicks.
An engine can win attention and send few clicks. An assistant built for writing, coding, or analysis keeps users inside the chat. An engine built as a search replacement pushes users to sources. That design choice decides how much of an engine's audience ever reaches your site.
The ratio between the two numbers tells you how "leaky" each product is. ChatGPT's referral share runs 1.45 times its visit share. Gemini's referral share runs at 0.43 times its visit share. Claude sits near one tenth. Perplexity's referral share runs 6.5 times its visit share, the highest click-out rate of any major engine.
Neither source is wrong. You need both: a buyer who reads your brand name inside Claude and never clicks still formed an opinion about you.
Claude's audience uses the product for work that ends inside the chat window. Similarweb's August 2026 reading put Claude at 9.3% of AI chatbot web visits, up from 1.9% twelve months earlier. That is close to a fivefold gain. StatCounter's September worldwide referral data puts Claude at 0.95%.
The B2B picture differs. Octane11's Q2 2026 report found Claude grew from 2.0% to 4.5% of AI referrals to B2B websites between March and June 2026, holding third place behind ChatGPT and Gemini. Claude sends business sites almost five times the share it sends the open web.
The enterprise footprint explains part of the gap. Anthropic told investors in August 2026 that annualized revenue had passed $65 billion. PwC committed to training and certifying 30,000 professionals on Claude, and TCS planned to roll Claude out to 50,000 employees across 56 countries. Those users research vendors, compare tools, and draft procurement memos. Few of those sessions produce a click, but many of them produce a vendor shortlist.
If you sell to enterprise buyers, Claude's visit share matters more to you than its referral share.
Perplexity has a small audience and a high click-out rate, and September data shows its referral share bouncing back. StatCounter's worldwide numbers put Perplexity at 7.91% of AI chatbot referrals in June 2026, 4.31% in August, and 5.85% in September. That is a 1.54-point recovery in one month after a 3.6-point slide over the summer.
Visit share tells a harsher story. Similarweb puts Perplexity at 0.9% of AI chatbot web visits in August 2026, behind DeepSeek (3.4%), Grok (2.4%), and Copilot (1.6%). In the United States, StatCounter's September referral data ranks Perplexity fourth at 3.02%, behind Copilot at 6.64%.
Perplexity's B2B number holds up. Octane11 measured it at 4.2% of AI referrals to B2B sites in June 2026, 0.3 points behind Claude. Perplexity users skew toward research queries, and research queries produce source clicks.
So Perplexity is a small engine that still sends a measurable share of qualified B2B visits. That profile calls for lower weight in your scorecard, not deletion.
Gemini's share flattened in Q3 after a year of fast gains. Similarweb's August 2026 data shows Gemini slipping from 27.8% of AI chatbot web visits three months earlier to 25.6%, while ChatGPT climbed back from 52.7% to 55.5%. Twelve months earlier, Gemini held 12.9%, so its share still doubled year over year.
StatCounter's referral data shows the same plateau. Gemini went from 7.94% of AI chatbot referrals in June to 10.9% in August, then held at 10.91% in September. In the US, Gemini sits at 13.26% of referrals, its strongest major market.
Octane11's B2B data shows more volatility. Gemini peaked at 17.3% of B2B AI referrals in April 2026, dropped to 9.3% in May, and recovered to 9.8% in June. Over the full September-to-June window, Gemini's B2B share still grew 130%.
Gemini's ceiling depends on distribution, not product preference. Google reported in August 2026 that the Gemini app passed 1 billion monthly users. Much of that activity runs through Android and Chrome, which Similarweb's web-only panel misses. Treat Gemini's web visit share as a floor.
ChatGPT sends about four out of five AI referrals to B2B websites, and Gemini, Claude, and Perplexity split most of the rest. Octane11's Q2 2026 report tracked ChatGPT's B2B share month by month: 86% in October 2025, a 69% low in April 2026, and a rebound to about 79% in June.
The same report puts AI referrals at under 2% of total B2B search sessions. Organic search still holds more than 96%. Reddit organic reached a record 1.9% of B2B search traffic in June 2026, close to the size of all AI engines combined.
Small volume does not mean small influence. Octane11 found that 1 in 11 closed-won B2B deals carried an AI search touch between August 2025 and June 2026, peaking at 14.5% of deals in February 2026. AI search sends you a thin stream of visits, and those visits show up in revenue.
StatCounter's broader view confirms the scale gap. In September 2026, AI chatbots accounted for 0.4% of referrals across search, social, and AI combined. Search held 91% and social held 8.61%. Weight engines against each other, but keep the channel in proportion.
Track AI referral share for your own brand → nobori.ai
Google's AI surfaces sit outside every chatbot market share table, and they reach more people than any chatbot. Google's June 2026 investor presentation put AI Overviews at more than 2.5 billion monthly users and AI Mode at more than 1 billion. Similarweb data reported by TechCrunch in July 2026 showed AI Overviews on 43% of US Google searches by May 2026, up from 15% a year earlier. AI Mode visits rose from 126 million in June 2025 to 279 million in May 2026.
None of this traffic appears as an AI referral in GA4. Clicks from AI Overviews and AI Mode arrive as Google organic. StatCounter's Google-versus-ChatGPT comparison for September 2026 gives Google 99.61% and ChatGPT 0.39% of combined referrals, which shows how far the search surface still outweighs the chatbots.
For tracking purposes, treat AI Overviews and AI Mode as their own tier. The two surfaces also disagree with each other. Recrawled's AI Shortlist 2026 study, run on September 4, 2026, found AI Overviews named the studied UK software companies 69 times across 240 shortlist questions, while AI Mode named them 44 times. Same company, same queries, different answers.
We covered how AI Overview citations translate into clicks in our AI Overview citation CTR analysis.
No, but you should stop giving it equal weight. Ross Hudgens, CEO of Siege Media, posted on LinkedIn in September 2026 that marketers should remove Perplexity from their LLM trackers, citing its shrinking market share. Greg Jarboe, writing in Search Engine Journal on September 9, 2026, agreed with the diagnosis and rejected the prescription. His proposal: de-weight Perplexity, keep monitoring it.
Jarboe sorted engines into three tiers. Tier one: ChatGPT and Gemini, each with its own score, plus Claude for B2B and enterprise audiences. Tier two: AI search embedded in existing ecosystems, meaning Google AI Overviews, AI Mode, and Microsoft Copilot for Microsoft 365-heavy buyers. Tier three: Perplexity, Grok, DeepSeek, and newer platforms, watched for spikes and growth.
Hudgens himself suggested a threshold: if Perplexity sends more than 5% of your AI referrals, keep it. StatCounter's worldwide September figure sits at 5.85%. Octane11's B2B figure sits at 4.2%. Most B2B sites will land near that line, which is why you should check your own analytics before you cut anything.
Microsoft says its first-party Copilot family has more than 150 million monthly active users. In the US, Copilot sends more AI referrals than Perplexity does.
Engine weighting can move a blended AI visibility score by more than 10 points with zero change in your actual visibility. Take the example Jarboe used: a brand cited in 40% of ChatGPT answers, 35% of Gemini answers, 30% of Claude answers, and 90% of Perplexity answers.
| Weighting method | Blended visibility score |
|---|---|
| Equal weight (25% each) | 48.75% |
| Weighted by StatCounter worldwide referrals (Sep 2026) | 42.3% |
| Weighted by Octane11 B2B referrals (Jun 2026) | 41.2% |
| Weighted by Similarweb web visits (Aug 2026) | 38.1% |
Nobori calculated the weighted figures by normalizing each source's shares across the four engines. The equal-weight score overstates this brand's position by 6.5 to 10.7 points, almost all of it from Perplexity. The brand looks strong on the engine with the smallest audience and average on the engines with the largest.
Run the same math in reverse and you see the risk. A brand that dominates Claude but scores low on Perplexity looks weaker than it is under equal weighting, even though Claude reaches ten times Perplexity's web audience.
Traffic share tells you where people are, not which engines your buyers trust for vendor decisions. Every engine answers the same question with a different set of brands. Recrawled's September 2026 study of 40 UK B2B software companies found that, among the 31 companies named on at least one platform, Gemini never named 13 of them (42%). Google AI Mode missed 11 (35%), AI Overviews missed 6 (19%), and ChatGPT missed 5 (16%).
Source preferences diverge too. The same study found YouTube in 25% of Google AI Overview answers about the companies and in none of the 360 ChatGPT answers. A video strategy can lift you on Google's surfaces and do nothing for ChatGPT.
This is why a single traffic-weighted score still hides problems. You can post a healthy blended number while you sit at zero on Gemini, the engine Google keeps embedding into Android, Chrome, and Workspace. Report the per-engine numbers next to the blended score, every time.
Our earlier analysis of AI citation overlap across engines shows how little the cited sources overlap between platforms.
Build your weights from three inputs: audience exposure, referral share, and your own conversion data. Start with this default and adjust after 30 days of your own analytics.
| Engine | Suggested B2B weight | Reasoning |
|---|---|---|
| ChatGPT | 40% | ~79% of B2B AI referrals and 55.5% of web visits |
| Google AI Overviews + AI Mode | 25% | 2.5B+ and 1B+ monthly users, invisible as AI traffic in GA4 |
| Gemini | 15% | 25.6% of web visits, 9.8% of B2B referrals, distribution through Android and Chrome |
| Claude | 12% | 9.3% of web visits, 4.5% of B2B referrals, enterprise-heavy audience |
| Perplexity | 5% | 0.9% of web visits, 4.2% of B2B referrals |
| Copilot | 3% | 2.5% of B2B referrals; raise for Microsoft 365-heavy buyers |
These weights are Nobori's starting point, not an industry standard. Shift weight toward Claude if your buyers are developers, analysts, or consultants. Shift toward Copilot if you sell into Microsoft-centric enterprises. Shift toward AI Overviews if your category already earns heavy organic traffic.
Revisit the weights each quarter. ChatGPT's B2B share moved 17 points between October 2025 and April 2026, then recovered 10 points by June.
Five changes take less than a day and fix most weighting errors.
Our crawl-to-refer ratio breakdown shows how much each engine takes from your site compared to what it sends back, a useful companion metric for these weights.
Expect the split to keep shifting by several points per quarter, with no single engine pulling away. ChatGPT regained share in Q3 after nine months of losses. Gemini's growth paused at a share twice its 2025 level. Claude kept compounding in visits and in B2B referrals. Perplexity stabilized at a small but measurable referral share.
Three forces could move the numbers in Q4. Google's new Gemini 4 models, which The Decoder reported as closing the gap with OpenAI and Anthropic, could restart Gemini's growth through Android and Chrome. OpenAI's mobile app and ChatGPT Work desktop client pull usage away from the web, which makes ChatGPT's web share look smaller than its true reach. Octane11 promised a Vol. 3 B2B report for late Q3 2026. When it lands, it will show whether Claude closed the gap on Gemini in B2B referrals.
Whatever happens, the measurement rule holds: weight engines by where your buyers are and what they do there. Recheck that weighting every quarter.
For a related look at how few brands each engine names per answer, read our analysis of the shrinking AI brand shortlist.
ChatGPT held 55.5% of AI chatbot web visits in August 2026 according to Similarweb, down from 73.3% a year earlier. StatCounter's September 2026 data gives ChatGPT 80.48% of AI chatbot referrals to websites worldwide and 75.17% in the US.
ChatGPT sent about 79% of AI referrals to B2B websites in June 2026, according to Octane11's analysis of 400 million B2B sessions. Gemini followed at 9.8%, Claude at 4.5%, Perplexity at 4.2%, and Copilot at 2.5%.
Claude users run tasks that finish inside the chat, such as writing, coding, and analysis, so fewer sessions end in a click to an outside site. Claude held 9.3% of web visits in August 2026 but 0.95% of worldwide referrals in September 2026.
Yes, at reduced weight. Perplexity held 0.9% of AI chatbot web visits in August 2026 but 5.85% of worldwide referrals in September and 4.2% of B2B AI referrals in June. Check your own analytics: if it sends more than 5% of your AI referrals, give it more weight.
No. StatCounter and Similarweb chatbot rankings exclude AI Overviews and AI Mode, and GA4 records their clicks as Google organic. Track them as a separate tier: Google reported more than 2.5 billion monthly AI Overview users and more than 1 billion AI Mode users in June 2026.
Get daily AI visibility alerts for your brand → nobori.ai
