速递｜OpenAI收购案后首秀：Windsurf推出SWE-1模型，向“自研AI玩家”的战略转身

Core Viewpoint - Windsurf has launched its first series of AI software engineering models, SWE-1, SWE-1-lite, and SWE-1-mini, optimized for the entire software engineering process rather than just coding [1][2]. Group 1: Product Launch and Features - The SWE-1 model is reported to perform comparably to Claude 3.5 Sonnet, GPT-4.1, and Gemini 2.5 Pro in internal programming benchmarks, although it still lags behind cutting-edge models like Claude 3.7 Sonnet in software engineering tasks [1]. - SWE-1-lite and SWE-1-mini will be available to all users for free or paid access, while SWE-1 will be limited to paid users [2]. - Windsurf claims that the service cost for the SWE-1 series will be lower than that of Claude 3.5 Sonnet [2]. - The company emphasizes that its models are designed to handle incomplete states, long-term tasks, and multi-platform scenarios, which are challenges for existing models [2]. Group 2: Company Strategy and Market Position - Windsurf aims to differentiate itself by focusing on the broader software engineering process rather than just coding, as highlighted by research director Nicholas Moy [2]. - The company is known for its "ambient programming" tools that allow software engineers to write and edit code through AI chatbots, competing with other startups like Cursor and Lovable [2]. - The launch of the SWE-1 series is described as a "proof of concept" for future AI model releases, indicating potential for further development [3].