en.Wedoany.com Reported - On June 9, the Guangnian Zhiwai (GN06) team under Meituan announced the official launch of the AI-native browser Tabbit 1.0. Tabbit uses the browser as an AI entry point, providing intelligent capabilities for web browsing, information retrieval, content processing, and complex task execution scenarios. The standard version is permanently free, with Windows and macOS versions available simultaneously. The mobile version has entered the testing phase, and the professional version is priced at 9.9 yuan per week.
Tabbit's core positioning is to expand the traditional browser from a "tool for opening web pages" to an "entry point for executing tasks." After users input their needs in the browser, the system can combine web page content, tab context, and built-in large model capabilities to complete multi-step operations across web pages and software, including workflows such as data retrieval, web page summarization, content rewriting, information organization, task decomposition, table extraction, email processing, subscription cleanup, and code reading. Compared to using an AI assistant as a chat window alongside a web page, Tabbit emphasizes enabling the AI to understand the pages the user is browsing and the relationships between multiple tabs, and through Agent mode, it connects search, reading, judgment, and execution into a continuous workflow. For users who frequently use browsers for office work, research, creation, and operational tasks, this type of AI-native browser is becoming an important form of large model implementation in personal productivity scenarios.
Tabbit launched its public beta on March 2 this year, and it has been approximately 100 days since then. During the public beta, the team maintained high-frequency iterations based on user feedback, cumulatively updating over a hundred features. The success rate of Agent tasks increased from the early 53.1% to 91.8%.
The launch of this 1.0 version also signifies that Meituan's exploration in the AI application layer is moving from model capability demonstration to a genuine product entry point. Browsers inherently possess high-frequency usage attributes, connecting search, content consumption, and office software, while also hosting web applications, plugin ecosystems, and account systems. By choosing the browser as its entry point, Tabbit can integrate multi-model invocation, web page understanding, automated execution, and personalized skills into a single interface, reducing the cost for users to switch between different tools. As AI Agent capabilities shift from "answering questions" to "completing tasks," the boundaries between browsers, operating systems, office suites, and search entry points are being redefined. Meituan's launch of Tabbit 1.0 reflects that Chinese internet companies are accelerating their efforts to compete for AI entry-level product positions.
The combination of a free standard version and a low-cost professional version is conducive to Tabbit rapidly expanding its user base in the early stages, and creating subsequent commercialization space through professional capabilities, task quotas, or advanced features. Future product competition will focus on task execution stability, privacy and data processing mechanisms, multi-model access quality, cross-application permission management, and mobile experience. If Agent capabilities can operate stably in more real-world office and content production scenarios, AI browsers have the potential to become an important carrier for large model applications to move from demonstrations to daily workflows.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
