TLDR: The fifth edition of Andreessen Horowitz’s Top 100 Gen AI Consumer Apps report indicates a stabilizing market for generative AI applications. While the growth of new web-based apps has slowed, the mobile sector is seeing more original entries following a crackdown on ‘ChatGPT copycats’. Google has significantly expanded its presence with four products on the list, including Gemini at number two, challenging ChatGPT’s dominance. Emerging trends like ‘vibe coding’ platforms are also gaining traction, and Chinese-developed apps are showing strong global export success.
The generative AI consumer application market is entering a phase of stabilization, according to the fifth edition of Andreessen Horowitz’s (A16Z) ‘Top 100 Gen AI Consumer Apps’ report, which analyzes two and a half years of data. The report, published on August 27, 2025, highlights a notable slowdown in new web-based applications, with only 11 newcomers compared to 17 in the March 2025 ranking. Conversely, the mobile app landscape saw 14 new entries, attributed to app store crackdowns on ‘ChatGPT copycats’ that have opened avenues for more original applications.
Google has made a significant impact on this edition of the list, with four of its products being independently ranked for the first time. Gemini, Google’s general Large Language Model (LLM) assistant, secured the second position behind ChatGPT on the web, capturing approximately 12% of ChatGPT’s web traffic. On mobile, Gemini’s performance is even stronger, achieving nearly half of ChatGPT’s monthly active users (MAUs), with a particularly robust presence on Android devices, accounting for almost 90% of its MAU base. Other Google products making their debut include AI Studio at #10, NotebookLM at #13, and Google Labs at #39. Google Labs, the company’s consumer-facing hub for AI experiments, experienced a more than 13% surge in traffic following the May 2025 launch of its video model, Veo 3.
In the broader competition among general LLM assistants, ChatGPT maintains its lead, but Google, X’s (formerly Twitter) Grok, and Meta AI are actively narrowing the gap. Grok, X’s assistant, ranked fourth on the web list and 23rd on mobile. Its mobile growth has been particularly impressive, transitioning from a ‘cold start’ at the end of 2024 to over 20 million MAUs. A significant 40% increase in mobile usage for Grok was observed in July 2025, coinciding with the release of its Grok 4 model on July 9 and the introduction of AI companion avatars, including the popular anime avatar ‘Ani’, on July 14. In contrast, Meta AI, which debuted in late May 2025, has shown more subdued growth, ranking 46th on the web and failing to make the mobile list, partly due to a June 2025 incident involving public feed posts.
The report also notes a flattening of mobile usage for DeepSeek and Claude, with DeepSeek experiencing a 22% drop from its peak. On the web, DeepSeek saw an even more dramatic decline of over 40% from its February 2025 peak, while Perplexity and Claude continued to grow.
China’s influence in the generative AI market is substantial, with three Chinese companies ranking in the web’s top 20: Quark (#9), Doubao (#12), and Kimi (#17). This is largely due to restricted access to non-Chinese LLM assistants like ChatGPT and Perplexity within China, necessitating local licensing and compliance with censorship. Furthermore, an estimated 22 of the top 50 mobile apps were developed in China and are now ‘exported’ globally, particularly in the photo and video editing categories, with companies like Meitu and Bytedance being prominent players. Chinese video models often benefit from a larger research focus and fewer intellectual property regulations.
An emerging trend highlighted in the report is ‘vibe coding’ platforms, such as Lovable (#22) and Replit, which allow users to generate apps quickly, often through natural language descriptions. These platforms demonstrate strong user retention, with U.S.-based users of one top platform showing over 100% revenue retention for several months post-signup. They also drive engagement with other AI products, as evidenced by the increased traffic to related services like Supabase.
Also Read:
- Chinese Gamers Embrace Generative AI in Gaming, Developers Rapidly Integrate Technology
- China’s Open-Source AI Models Rapidly Closing Gap with Global Proprietary Leaders
The report identifies 14 ‘All Stars’ – companies that have consistently appeared on all five editions of the web top 50 list over the past two and a half years. These include diverse applications spanning general assistance (ChatGPT, Perplexity, Poe), companionship (Character AI), image generation (Midjourney, Leonardo), image and video editing (Veed, Cutout), voice generation (Eleven Labs), productivity tools (Photoroom, Gamma, Quillbot), and model hosting (Civitai, HuggingFace). Notably, five of these All Stars utilize proprietary models, seven rely on API-available or open-source models, and two function as model aggregators. These consistent performers originate from just five countries: the U.S., UK, Australia, China, and France.
