The Top 100
Gen AI
Consumer Apps
In just six months, the consumer AI landscape has been redrawn. Some products surged, others stalled, and a few unexpected players rewrote the leaderboard overnight. Deepseek rocketed from obscurity to a leading ChatGPT challenger. AI video models advanced from experimental to fairly dependable (at least for short clips!). And so-called “vibecoding” is changing who can create with AI, not just who can use it. The competition is tighter, the stakes are higher, and the winners aren’t just launching, they’re sticking.
We turned to the data to answer: Which AI apps are people actively using? What’s actually making money, beyond being popular? And which tools are moving beyond curiosity-driven dabbling to become daily staples?
This is the fourth installment of the Top 100 Gen AI Consumer Apps, our bi-annual ranking of the top 50 AI-first web products (by unique monthly visits, per Similarweb) and top 50 AI-first mobile apps (by monthly active users, per Sensor Tower). Since our last report in August 2024, 17 new companies have entered the rankings of top AI-first web products. 
A methodology note: Products that have added significant generative AI features but are not AI-native, such as Canva and Notion, are not included here. In a change from prior editions, this means we’ve removed legacy photo editing products, including Pixlr, Fotor, and PicsArt.
In addition, for this edition we’re introducing the “Brink List”: the next 10 companies — five on web, five on mobile — that just missed the cut-off. The AI world moves fast, and we’re curious to see which (if any!) of these products make the top 100 ranking next time. 
Beyond these rankings, our analysis uncovers telling signals about where AI is gaining traction and emerging consumer tech behaviors.
Here are some of our top takeaways:
When ChatGPT launched in research preview in November 2022, it became the fastest-growing consumer app ever, reaching 100 million users in just two months. However, usage soon plateaued, with global monthly visits nearly flat between March 2023 and April 2024.
More recently, growth has come roaring back. It took nine months for ChatGPT to grow from 100 million weekly active users in November 2023 to 200 million in August 2024. Typically, growth becomes more challenging at scale, but in this case it took less than six months to double again — reaching 400 million weekly active users by mid-February 2025. 
What’s driving this growth? The long initial plateau stemmed from ChatGPT’s early novelty factor; many consumers found it intriguing, but lacked compelling daily use cases. However, as OpenAI has introduced more advanced models and capabilities into ChatGPT, usage has risen accordingly — both among existing users and a wave of new adopters.
The biggest spikes align with product milestones. The April-May 2024 surge? That was GPT-4o, introducing multimodal capabilities. Users could talk to ChatGPT in real-time, show it images, and get instant, intelligent responses — from snapping a photo of a tricky math problem to having an AI-powered brainstorming session.
The July-August 2024 jump followed the launch of Advanced Voice Mode, bringing near-human fluency to AI conversations. And the September-October 2024 boost? That was the debut of the o1 model series, which leveled up reasoning and problem-solving. 
ChatGPT’s mobile data tells a more consistent growth story. Since its launch in May 2023, monthly active users have steadily increased by 5 to 15% every month over the past year, as web users also adopted the app. According to Sensor Tower estimates, out of ChatGPT’s 400 million weekly active users, 175 million now use the mobile app.
DeepSeek made a big splash in more ways than one. Though its public chatbot only launched on January 20, it racked up enough traffic in just 10 days to rank #2 globally among all AI products for the month.
DeepSeek was developed by the Chinese hedge fund High-Flyer, and the plurality of usage (21%) in January 2025 came from China, where ChatGPT is banned. Other top countries for DeepSeek include the U.S. (9% of traffic) and India (8%). DeepSeek has been banned or restricted in some countries, including South Korea, Australia, and Taiwan, as well as on government devices in several U.S. states.
DeepSeek’s launch significantly outpaced most other general assistant LLMs. The company reported 1 million users in 14 days — slower than ChatGPT’s 5-day mark — but then surged past 10 million users in just 20 days, beating ChatGPT’s 40-day milestone.
On mobile, DeepSeek launched on January 25 and hit the #14 spot for monthly active users in just five days. By February, it had jumped to #2 on mobile, capturing 15% of ChatGPT’s mobile user base.
These are not just “empty downloads.” Per Sensor Tower data, DeepSeek users are slightly more engaged on mobile than users of Perplexity and Claude, based on both sessions per user and minutes per user in the average week. However, engagement still significantly trails behind ChatGPT. 
DeepSeek’s growth among the tech and research community was driven by its strong performance on reasoning benchmarks and its headline-grabbing claim of achieving those results with just $5.6 million in training costs — a fraction of what competitors have spent.
Its claim about training costs sparked broader media attention, generating mainstream discussions of a “Sputnik moment” for AI. According to Google Trends, global search interest for DeepSeek matched ChatGPT on January 27, and even surpassed it in the U.S.!
AI video has been on the brink of true usability (with reliable outputs) for the past 18 months, but the last six months finally delivered major progress in quality and controllability. Three new companies — Hailuo (#12), Kling AI (#17), and Sora (#23) — debuted on the web rankings, joining InVideo (#37). Runway and Krea (which allows users to generate videos across models) made the Brink List.
Hailuo (maker of the MiniMax model series) and Kling are both Chinese video models — Kling launched in June 2024, followed by Hailuo in September. As of January 2025, both surpassed Sora in monthly visits. While Sora was first previewed in February 2024, it wasn’t released until December 2024.
We are beginning to see fragmentation across video model providers. They’re diverging in both output style and user interface, offering creators more specialization. Sora is known as a versatile, all-around video generator, Hailuo excels in strong prompt adherence, and Kling offers extra features like camera movement control and lip sync.
AI-powered video editing remains a top consumer AI use case, streamlining tasks like intelligent clipping, captioning, and other time-consuming processes into “one click” solutions. Veed (#36) and Clipchamp (#45) made the web rankings, while mobile had the largest representation from hybrid video-photo editing apps, led by B612 (#12), VivaCut (#15), and Filmora (#19). In terms of revenue, Splice, Captions, and Videoleap were the top-performing apps (more on that below).
We expect even more movement here in the coming months, with models like Google’s Veo 2 (not yet publicly available) poised to raise the bar, based on feedback from early testers. Priced at $0.50 per second of video, however, Veo 2 is likely geared more toward commercial use than mass consumer adoption.
It’s the era of the builder — and AI-powered tools are rising to meet demand. Two categories of products have exploded in the last six months, targeting different audiences: (1) agentic IDEs (integrated development environments) designed for developers; and (2) text-to-web app platforms accessible to both technical and non-technical users (deemed “vibecoders” by Andrej Karpathy).
Agentic IDEs like Cursor act as co-pilots for developers, with features like bug checking, autocomplete, and full code generation. Cursor debuted at #41 on our web ranking — the company recently announced that it serves hundreds of thousands of developers.
By contrast, text-to-web app products don’t require users to be technical. Anyone can input a text prompt and get a usable website or web application, which can feel like magic. These products are emerging now for a variety of reasons: models are finally able to generate executable code, web frameworks have matured, and platforms can take advantage of libraries and SDKs from companies like Resend, Clerk, and Supabase.
Like agentic IDEs, leaders in this space are growing fast. Bolt debuted at #48 on the list, while Lovable landed on the Brink List. Bolt reported reaching $20 million in annualized revenue and 2 million registered users within its first two months, while Lovable reported growing to $17 million in annualized revenue in its first three months.
In terms of monthly unique visitors (not visits) on web, the top two agentic IDEs are currently outpacing the top two text-to-web app builders, though both categories are growing at a similar rate. We’re interested to see how this evolves. Agentic IDEs require users to be technical, but every developer can (theoretically!) find value from using one on a near-daily basis. In contrast, text-to-web app platforms cater to the 99% of consumers who aren’t coders, but require users to have an idea of what they want to build. 
There is some overlap between the user bases of these two product types. Technical users might use text-to-web app platforms for prototyping or generating code they can download and refine. According to Similarweb data, 23% of Bolt’s unique visitors in January 2025 also visited Cursor.
Our rankings of mobile apps are based on global monthly active users (MAUs), as measured by Sensor Tower. However, the most popular apps don’t always generate the most revenue.
To make our list, an app needed more than 8 million MAUs, yet many with far fewer users achieved significantly higher conversion rates and revenue per user. ChatGPT was the clear leader on both metrics. But in our analysis of the top 50 mobile apps by MAUs versus those generating the most monthly revenue (from mobile-originated subscriptions), only 40% overlapped.
Of the 80 apps that ranked in the top 50 for either revenue or usage, many with lower usage proved to be the most effective at monetization — at least through mobile subscriptions, as ad revenue wasn’t accounted for. 
Some product categories were represented on both lists, but by different companies. Photo and video editors composed 20% of the revenue list and 24% of the MAU list. However, the top three dedicated video editors by usage — VivaCut (#15), Filmora (#19), and Beat.ly (#36) — differed from the top three by revenue: Splice, Captions, and Videoleap. The apps with the broadest consumer appeal are not always the apps with the highest-grade prosumer, or even enterprise-level, features that more serious users will pay for.
Some product categories ranked for revenue, but had no apps on the usage list, including plant identification (PictureThis, PlantID), nutrition (Cal AI, Fastic), language learning (Speak, Learna, Loora), music (Moises, Suno), and dictation (Otter, PLAUD). These apps serve specific goals or skills, making them less universal, but more likely to attract users who are willing to pay for their specialized value.
Interestingly, ChatGPT “copycat” apps represented 12% of both the mobile usage and revenue lists. We first covered these in our initial mobile ranking, noting their strategy of mimicking ChatGPT’s name and logo to suggest they offer the same premium model at a lower price. They frequently tweak their listings as a way to avoid app store enforcements and copyright claims.
* * *
AI-native products are scaling faster and engaging users more deeply than ever. Though things are moving fast, there’s still a lot of ground left to break. AI will underpin category-defining companies in the years ahead, but the race to shape its future is already well underway — gaining momentum, mainstream adoption, and real revenue.
