Other languages：繁體中文｜日本語｜한국어

This article reflects information available as of August 2026. User counts and product features may continue to change.

In August 2026, Google announced that the Gemini App had surpassed 1 billion monthly users, making it the fastest-growing product in Google’s history. OpenAI’s latest publicly confirmed ChatGPT metric, meanwhile, has reached 1 billion weekly users. Both products have crossed the 1 billion mark, but monthly active users and weekly active users are fundamentally different metrics. Add Google’s distribution advantages through existing platforms such as Android, and the headline number alone tells us very little about which product has stronger engagement.

This was not the first time Gemini had rapidly approached the milestone. In November 2025, the Gemini App had around 650 million monthly users. That increased to 750 million in February 2026, more than 900 million by Google I/O in May, and 950 million during the July earnings call before officially crossing 1 billion in August. The Verge noted that Gemini is now the 14th Google product to reach a billion-user scale. For people who actually use ChatGPT, Gemini, Claude, and other AI tools every day, these numbers are useful for understanding market position, but they are not a good way to decide which product is better for a particular workflow.

Gemini and ChatGPT Have Both Passed 1 Billion Users, but the Numbers Measure Three Different Things

Google’s 1 Billion Means Monthly Gemini App Users, While OpenAI Has Confirmed 1 Billion Weekly ChatGPT Users

Google’s official statement on August 11 was clear: the Gemini App had surpassed 1 billion monthly users. Google did not disclose the full methodology behind that figure, such as the minimum activity threshold required to count as a user or exactly how users are deduplicated across devices. It is therefore safest to understand the number as Google’s own monthly-user metric for the Gemini App.

OpenAI’s situation is slightly different. A usage report published on August 6 said that “more than 1 billion people are using ChatGPT,” but that article itself did not label the figure as WAU. However, another official OpenAI product announcement explicitly states that “Every week, 1 billion people turn to ChatGPT,” and an OpenAI spokesperson also told The Verge that ChatGPT had crossed 1 billion weekly users in July.

What can therefore be confirmed at this point is that ChatGPT has reached 1 billion weekly users, while Gemini’s official figure is 1 billion monthly users.

Metric	Gemini	ChatGPT
Latest company figure	1 billion+	1 billion
Time period	Monthly users	Weekly users
Announcement timing	August 11, 2026	Officially confirmed in August 2026
Product scope	Gemini App	ChatGPT
Can they be directly compared?	Not directly	Not directly

Weekly activity generally represents a stricter engagement threshold than monthly activity because the same user must appear within a much shorter seven-day period. But without both companies publishing their complete measurement methodology, “1 billion WAU versus 1 billion MAU” still cannot be used to calculate the actual engagement gap between the two products.

Sensor Tower Estimated ChatGPT Mobile Had Already Passed 1 Billion MAU in May—but That Is a Third Measurement Standard

If third-party data is included, ChatGPT actually appeared to cross the 1 billion monthly-active-user threshold earlier.

Sensor Tower’s State of AI 2026 estimated that ChatGPT became the fastest mobile App in history to reach 1 billion monthly active users in May 2026. Reuters also reported that figure in June.

However, this number still cannot be directly compared with Google’s 1 billion figure.

Sensor Tower was measuring the ChatGPT mobile app, while Google’s number refers to the broader Gemini App product across web and different devices. The former is a third-party estimate, while the latter is based on Google’s internal data. Both the source and scope are different.

So it is fair to write that “ChatGPT reached an estimated 1 billion MAU three months before Gemini,” but only with a clear qualifier: this refers to Sensor Tower’s third-party estimate for the ChatGPT mobile app, not an official OpenAI MAU figure calculated using the same methodology as Google.

Does Gemini’s 1 Billion Include Google Search? Apparently Not

Google’s Figure Refers to the Gemini App, Not Total Exposure to Google AI

Gemini has already been integrated into Search, Chrome, Workspace, Android, and other Google products, but this 1 billion figure does not appear to simply add together everyone who has encountered Google AI somewhere in the ecosystem.

The Verge confirmed with Google that the figure appears to correspond directly to the Gemini App. AI Mode in Google Search has its own separate usage statistics and had already surpassed 1 billion monthly users by May 2026.

This distinction matters.

If everyone who saw an AI-generated answer in Search were included in Gemini’s 1 billion, the meaning of the number would be very different. The more reasonable interpretation is that the Gemini App itself has now reached a billion-user monthly scale, while Search AI Mode represents another separate Google AI entry point that has also passed 1 billion users.

Android Distribution Is Clearly an Advantage for Gemini, but We Cannot Calculate How Many Users Came Only Because It Was the Default

Google still has an obvious distribution advantage.

Gemini is deeply integrated into Android, and Google has continued shifting phone functions previously handled by Google Assistant over to Gemini. In this latest announcement, Google also said Gemini can now perform actions such as ride-hailing and reservations across more than 40 popular Apps.

That means Gemini does not need to rely entirely on users independently discovering and downloading an unfamiliar AI App. This is a very different position from an independent AI company.

However, the article should not claim that “a significant portion of the billion users came from default distribution rather than active downloads,” because Google has not published a breakdown showing how much each acquisition source contributes.

What can be confirmed is that Android preinstallation and system-level integration reduce the friction of trying Gemini. What cannot be confirmed is how many monthly users are counted primarily because Gemini was available as a default or deeply integrated entry point.

What Matters More Than the 1 Billion Figure Is How Gemini Is Actually Being Used

63% of Gemini Users Speak to It Directly

Among the behavioral statistics Google published, one of the most notable is that 63% of Gemini users interact with the product using voice, while the number of users who rely only on voice is also increasing. Busy parents are reportedly 43% more likely than other groups to use voice for everyday tasks.

These numbers suggest that Gemini’s current positioning is increasingly different from that of an early text-only Chatbot website.

Android integration, Gemini Live, mobile cameras, and cross-App actions make Gemini easier to use while walking, driving, doing housework, or handling tasks when the user’s hands are occupied.

For people who mainly use AI to write long documents, organize research, or code, this direction may not immediately change their workflow. For users who already handle much of their daily life through a phone, however, the voice interface may matter much more.

“One in Five Uses a Camera or Screen Sharing” Applies Only to Gemini Live, Not All Gemini Conversations

An earlier version of this article needs correction here.

Google’s actual statement was that one in five Gemini Live interactions involves more than voice. Users also activate the live camera or share their screen so Gemini can see what is happening and help with the problem.

This does not mean that 20% of all Gemini Prompts use the camera.

Google specifically highlighted DIY users and students as groups making frequent use of these features. Real-world scenarios could include pointing the phone camera at a device and asking which part is malfunctioning, or sharing the current screen and asking how to complete a task.

These behaviors also help explain why Google increasingly presents Gemini not simply as a text chatbot, but as a multimodal Assistant.

Gemini Now Generates More Than 150 Million Images Per Day

Google also said Gemini currently generates more than 150 million images every day.

That volume suggests image generation has moved beyond being an occasional novelty and is now a high-frequency use case. Google specifically noted that small businesses use the feature for marketing materials and other content.

The 150-million-image figure alone does not tell us how many of those images are eventually published, nor can it directly explain the growth of AI-detection tools.

But it does provide a useful sense of scale: one product alone can now generate more than 100 million AI-created images per day. As the cost of producing synthetic content falls, it is not surprising that platforms are increasingly focusing on provenance labels, content quality, and management of generated material.

Gemini Has More Than 100 Million Users on Apple Devices—What Does That Tell Us?

Google said Gemini has more than 100 million active users on iOS, while macOS power users send Prompts at roughly twice the rate of users on other interfaces.

Both numbers were included in Google’s latest official announcement.

The iOS figure is particularly interesting because Gemini is not the system-default AI Assistant on the iPhone. Users generally need to acquire the Gemini App separately or actively enter the service.

At minimum, this shows that Gemini’s growth cannot be explained entirely by Android distribution.

However, it would also be wrong to interpret those 100 million users as pure product preference completely separate from the Google ecosystem. Gmail, Google Maps, YouTube, and other Google products already have enormous audiences on iOS.

The macOS figure also needs to be read carefully because of Google’s qualifier: macOS power users.

Google did not say all macOS users send twice as many Prompts as users on other platforms. Power users are already a high-frequency segment, so this statistic is more useful for understanding the behavior of a specific group of heavy users than for claiming Mac users overall are twice as engaged with Gemini.

After Reaching 1 Billion Users, Google Still Has to Answer How AI Becomes a Business

Google Has Not Disclosed How Many Consumer Gemini Users Actually Pay

Google currently offers several paid AI tiers, including Google AI Plus, Pro, and Ultra. In the United States, AI Plus costs $7.99 per month, while Google also introduced a new premium AI Ultra plan at I/O 2026.

The unanswered question is how many Gemini App users actually pay for these plans.

Alphabet has not separately disclosed a consumer Gemini subscriber count.

In its first-quarter 2026 earnings, Alphabet reported 350 million paid subscriptions across its broader consumer services and said it was the strongest quarter so far for consumer AI plans, driven in part by adoption of the Gemini App.

But that 350 million figure also includes subscriptions across YouTube, Google One, and other services. It cannot be interpreted as 350 million paid Gemini users.

What is known is that Gemini has surpassed 1 billion users and that Google says consumer AI subscriptions are growing. What remains unknown is how many of those billion users directly pay for Gemini and how much revenue each user generates on average.

At the Same Time, Alphabet Raised 2026 Capital Expenditure Guidance to $195–205 Billion

When Alphabet reported second-quarter earnings in July, it raised its expected 2026 capital expenditure from a previous range of $180–190 billion to $195–205 billion.

The company said the increase was largely driven by the need to accelerate delivery of computing capacity to meet demand, while also warning that infrastructure costs related to data centers, energy, and depreciation would continue increasing.

Those investments are not all for the Gemini App. Alphabet also needs to support Search, Cloud, TPUs, and other AI products, so dividing $200 billion by Gemini’s user count would not produce a meaningful cost-per-user figure.

What it does show is that acquiring AI users at billion-person scale and turning that usage into a financially sustainable business model are two separate problems.

How Should AI Services Charge? The BBC’s “Tokenomics” Problem Fits Directly into This Debate

A BBC report from early August described the current challenge as Tokenomics: businesses buying AI services struggle to control spending, while providers are still trying to determine how these products should be priced.

AI does not behave like traditional seat-based SaaS.

One employee may ask only a few simple questions in a day, while another may run an Agent for hours, load large amounts of context, and perform repeated rounds of reasoning. The compute cost generated by those two users can differ dramatically.

This issue resembles recent debates over Notion AI usage limits and Custom Agents Credits, but it would be inaccurate to write that “Notion limits usage because upstream AI costs are impossible to calculate.” There is currently no evidence establishing that direct causal relationship.

A more defensible connection is that the entire AI software industry is dealing with the same economic problem: usage is difficult to predict, but consumers are accustomed to fixed monthly subscriptions. Providers therefore need to balance unlimited usage, caps, Credits, Token-based pricing, and different plan tiers.

Gemini 3.5 Pro Delays and DeepMind Leadership Changes Add Two More Variables Beyond the 1 Billion Users

Gemini 3.5 Pro Was Originally Expected in June but Had Still Not Been Publicly Released by August

At I/O 2026 in May, Google explicitly said it was developing Gemini 3.5 Pro and expected to release it “next month,” which pointed to June.

By July, Google had already released Gemini 3.6 Flash and 3.5 Flash-Lite, but the official Gemini API Release Notes still did not show a formal Gemini 3.5 Pro launch.

Reuters reported in July that the Gemini 3.5 Pro delay was related to the model not meeting internal performance targets, particularly in Coding, an area that has become increasingly important in enterprise AI competition.

During Alphabet’s second-quarter earnings call, Sundar Pichai also acknowledged that Coding and Agentic Coding remained areas where Google needed to improve, while saying the company had already begun training a larger Gemini 4 model.

This does not mean Gemini’s billion-user consumer base will necessarily decline because of the delay.

Consumer reach and Frontier Model rankings are not the same thing. But for teams considering whether to build long-term API, developer-tool, or production workflows around a specific model provider, release cadence matters more directly than App user counts.

Demis Hassabis Moved Away from Day-to-Day Management, While Jeff Dean Left Google After 27 Years

Google’s AI leadership also changed significantly in August.

Demis Hassabis stopped serving as Google DeepMind CEO and moved into the roles of Google DeepMind Chair and Alphabet Chief Scientist, shifting more of his focus toward research and science. Former CTO Koray Kavukcuoglu took on more responsibility for daily operations and model development.

Jeff Dean, meanwhile, left Google after 27 years and co-founded Discovery Loop with other researchers.

An earlier version of this article described this as “multiple senior research leaders leaving, including Jeff Dean, who had shaped Google engineering for nearly three decades.” That direction was broadly correct, but Hassabis and Dean should not be grouped under the same category of “departure.”

Hassabis remains at Alphabet but is no longer responsible for DeepMind’s day-to-day management. Dean actually left Google.

These leadership changes and the Gemini 3.5 Pro delay are both worth watching, but there is currently not enough evidence to show that either has already affected Gemini App user growth.

What is clearer is that two things are happening simultaneously: Gemini App grew from 750 million to more than 1 billion monthly users within roughly half a year, while Google is restructuring parts of its model-development organization and the release schedule for its next flagship model has not followed the original timeline exactly.

After Gemini Passes 1 Billion Users, Choosing an AI Tool Should Still Start with Your Own Workflow

Do Not Use the 1 Billion Figure to Decide Which AI Is “Better”

Gemini and ChatGPT have both reached billion-user scale, but Google reports MAU, OpenAI has confirmed WAU, and Sensor Tower separately estimated ChatGPT mobile-app MAU.

These figures are useful for understanding how large the market has become. They are not a model-quality leaderboard.

What matters in everyday use is still where your data lives, which tools you need, whether you frequently use voice or cameras, whether you need Agents, and whether your workflow depends on Gmail, Docs, Notion, GitHub, or other systems.

Google’s Android and Workspace integrations may be highly convenient for some teams. ChatGPT or another tool may fit a different workflow better.

One billion users cannot answer those questions for you.

Gemini’s Voice and Live Numbers Are More Useful for Understanding What Google Is Optimizing

The fact that 63% of users speak to Gemini, that one in five Gemini Live interactions includes a camera or screen-sharing component, and that Android can use Gemini to act across more than 40 Apps tells us more about Google’s product direction than a single MAU figure.

If your main needs involve mobile voice, working directly from what is on the screen, or real-time assistance while moving around, this product direction is closely aligned with that workflow.

If your main work involves long-document research, software development, or large-scale text processing, then model quality, context handling, tool integrations, and cost need to be evaluated separately.

User count tells you that a lot of people are touching the product. It does not tell you that every type of task belongs there.

Beyond Price, Pay Attention to Usage Limits and Model Update Policies

The AI subscription market is still not fully settled.

API prices can fall, consumer plans can introduce new usage limits, and premium tiers can change as infrastructure costs and competition shift.

Alphabet is now operating at an annual capital-expenditure level approaching $200 billion, while the enterprise market discussed by the BBC is still grappling with unpredictable Token consumption.

So if a long-term workflow depends heavily on one AI provider, it may be more important to track how usage is measured, whether data can be moved elsewhere, whether alternative models are supported, and what switching costs exist than whether the subscription is a few dollars cheaper this month.

Gemini App surpassing 1 billion monthly users is unquestionably a major product milestone. Its rise from 650 million in November 2025 to more than 1 billion in August 2026 also helps explain why Google is integrating Gemini into an increasing number of products.

But “1 billion people use Gemini every month” primarily answers a question about reach.

It does not tell us how many times those users interact with the product each month, how many actively chose Gemini over alternatives, how many pay for it, or which AI performs best for writing, programming, research, voice, or Agent workflows.

When these metrics are not separated carefully, market scale can easily be mistaken for product preference.

For actual users, whether the tool can complete the work in front of them and whether its cost is predictable are still more useful questions than the billion-user number at the top of the leaderboard.

FAQ

SUPPORT FENGNIII

喜歡這篇文章嗎？

如果這篇內容對你有幫助，可以透過小額贊助支持本站持續整理更多日文、韓文、旅行與數位工具內容。

小額支持本站

付款將由藍新金流安全處理