On July 22, 2026, on Alphabet's second-quarter earnings call, Sundar Pichai said the Gemini app now has 950 million monthly active users, with daily active users tripling over the past year. That puts Google about 50 million users away from the billion-user mark ChatGPT crossed earlier this year — and the tripling of daily use is arguably the more telling number, because daily use is a much harder thing to win than a one-time visit. Two AI assistants are now operating at a scale almost nothing else in consumer software reaches. The interesting question is not who is ahead by 50 million — it is what a number like that actually tells you, and what it very deliberately does not.
Because here is the thing about 950 million: Google can put Gemini in front of you without you ever choosing it. It is in Search, in Android, in Chrome, in Workspace. ChatGPT's billion users mostly went and got it. Those are different kinds of numbers wearing the same units, and understanding the difference is the whole story of the AI assistant race in 2026.
Quick summary: Gemini reached 950 million monthly active users, disclosed on Alphabet's Q2 2026 earnings call on July 22, with Pichai saying daily active users tripled year over year. Google's AI Mode in Search passed 1 billion. Alphabet reported quarterly revenue of $119.8 billion, up 24 percent, with Google Cloud up 82 percent to $24.8 billion. Gemini models now process about 22 billion API tokens per minute, up from 16 billion the previous quarter. But distribution is not preference: Google reaches users through Search, Android and Workspace, while ChatGPT's audience largely sought it out. Neither number measures which assistant answers your question better.
The Numbers Google Actually Reported
It is worth separating the AI headline from the business underneath it, because the second explains the first. Alphabet reported quarterly revenue of $119.8 billion, up 24 percent year over year. Google Cloud grew 82 percent to $24.8 billion, which the company attributes largely to enterprise demand for AI infrastructure and Gemini tooling. Roughly 90 percent of the Fortune 100 are now using Gemini Enterprise. And the raw usage figure that matters most for infrastructure planning: Gemini models are processing about 22 billion API tokens per minute, up from 16 billion the prior quarter. Sources: Google; 9to5Google; CNBC.
| Metric | Q2 2026 | What it signals |
|---|---|---|
| Gemini app monthly users | 950 million | Nearing ChatGPT's billion; up from about 750M in Q4 2025 |
| Gemini app daily users | Tripled year over year | The harder-won number — daily habit, not a one-time visit |
| AI Mode in Search | Over 1 billion | Larger than the app — most people meet Gemini in Search |
| Alphabet revenue | $119.8 billion | Up 24 percent year over year |
| Google Cloud revenue | $24.8 billion | Up 82 percent, driven by AI demand |
| API tokens per minute | About 22 billion | Up from 16 billion last quarter |
| Fortune 100 on Gemini Enterprise | About 90 percent | Enterprise adoption is not the bottleneck |
Why 950 Million Is Not the Same as 950 Million
The single most important line in that table is the second one: AI Mode in Search passed a billion, which is more than the Gemini app itself. That tells you where most Gemini exposure comes from. You searched for something, Google generated an AI answer, and you were counted. You did not download anything, you did not sign up, you may not have thought of it as using an AI assistant at all.
Compare that to how a typical ChatGPT user arrives. They heard about it, went to a website or an app store, created an account, and came back. That is a much higher-effort action, and it signals something a passive impression does not: intent. This is not a knock on Google — owning the front door of the internet is a genuine and enormous advantage, and it is exactly why Gemini scaled this fast. But it does mean the two billion-ish numbers are measuring different behaviours, and anyone using them to declare a winner is comparing things that are not alike.
What Google Shipped the Day Before
The user milestone landed a day after a notable model release. On July 21, Google shipped Gemini 3.6 Flash, which it says uses about 17 percent fewer output tokens than its predecessor and costs less than 3.5 Flash, alongside 3.5 Flash-Lite and a restricted 3.5 Flash Cyber limited to governments and trusted partners. Notably absent: Gemini 3.5 Pro, which Google said is not ready. The company also confirmed it has begun its largest pretraining run so far, for Gemini 4. Sources: TechCrunch; Droid Life.
That release pattern is informative. Shipping cheaper, more efficient Flash models while holding back the flagship Pro model suggests Google is currently optimising for cost per query at enormous volume rather than for topping benchmark charts. When you are serving a billion people in Search, a token reduction like that is worth more than a leaderboard position.
So Which One Should You Actually Use?
User counts are a terrible basis for that decision, and here is the honest framing instead. Gemini has the strongest position if you live inside Google's products — it reaches Gmail, Calendar, Docs and Search natively, and for many people that context is worth more than raw model quality. ChatGPT retains the deepest third-party ecosystem and the most mature voice and agent tooling. Claude tends to be preferred for long-form writing and code by people who do that work daily. None of that is settled by who has more monthly actives.
- Pick on the task you actually do most, not the market-share headline. The gap between these assistants on any single question is usually smaller than the gap between a good prompt and a lazy one.
- If you already live in Gmail, Docs and Calendar, Gemini's integration is a real and underrated advantage — it can see context the others cannot.
- If you want the widest set of third-party integrations and the most developed voice mode, ChatGPT is still ahead on ecosystem maturity.
- Free tiers differ more than the models do. Check what each gives you before paying anything.
- Treat any 'X is winning' article built purely on user counts with suspicion, including the ones agreeing with your current preference.
The Part Worth Watching Next
Google confirming it has started its largest pretraining run ever, for Gemini 4, is the more consequential piece of news buried under the user milestone. Pretraining runs of that size take months and cost enormous sums, and companies do not announce them casually. Combined with holding 3.5 Pro back as not ready, the read is that Google is choosing to skip an incremental flagship in favour of a larger jump. Whether that pays off is the story of the next two quarters, and it will matter far more to which assistant you use in 2027 than today's 50-million-user gap.
There is a practical takeaway underneath all of this. The assistants are converging on quality faster than they are differentiating, which means the cost of picking wrong is falling and the cost of being locked in is rising. Being able to put the same question to several current models and compare the answers is more useful than committing to one vendor's roadmap. That is the entire premise of LumiChats — many current models under one login at a pay-per-day price, so you can judge them on your own work rather than on an earnings-call number.
01How many users does Gemini have in 2026?
The Gemini app reached 950 million monthly active users, disclosed by Sundar Pichai on Alphabet's Q2 2026 earnings call on July 22, 2026. Google said its daily active users tripled over the past year; the monthly figure rose from about 750 million in late 2025. Separately, AI Mode in Google Search passed 1 billion.
02Is Gemini bigger than ChatGPT now?
Not quite, and the comparison is imperfect. Gemini's app is at 950 million monthly users; ChatGPT crossed 1 billion earlier in 2026. But Google reaches many of those users through Search and Android rather than by deliberate choice, so the two numbers measure somewhat different behaviour.
03What did Google release alongside this?
On July 21, 2026 Google shipped Gemini 3.6 Flash — which it says uses about 17 percent fewer output tokens than its predecessor — plus 3.5 Flash-Lite and a restricted 3.5 Flash Cyber. Gemini 3.5 Pro was held back as not ready, and Google confirmed it has begun its largest pretraining run yet for Gemini 4.
04Does a bigger user base mean a better AI?
No. User counts measure distribution and defaults, not answer quality. Google places Gemini inside Search, Android, Chrome and Workspace, which drives enormous exposure independent of whether users prefer it. Judge an assistant on your own tasks.
05How much revenue is Google making from AI?
Alphabet reported Q2 2026 revenue of $119.8 billion, up 24 percent year over year, with Google Cloud up 82 percent to $24.8 billion, attributed largely to enterprise AI demand. About 90 percent of the Fortune 100 now use Gemini Enterprise.
06Which AI assistant should I pay for?
It depends on where you work. Gemini wins if you live in Gmail, Docs and Calendar. ChatGPT has the most mature ecosystem and voice tooling. Claude is widely preferred for long-form writing and coding. Compare free tiers first — they differ more than the models do.
The bottom line: Gemini at 950 million monthly users is a genuinely staggering distribution achievement, and it says more about Google owning Search and Android than about Gemini beating anyone on quality. The number worth watching is not the gap to a billion — it is that Google is skipping an incremental flagship to bet on Gemini 4. That decision will shape which assistant is actually best a year from now, and no earnings call can tell you that yet.
